Witness Index

Vulnerabilities Multiple significant regulatory accuracy problems in this partial direct examination: (1) CONGREGATE REQUIREMENT: Honer testified that meals must be eaten in congregate settings in both CACFP and SFSP programs. This was a pre-COVID requirement. USDA issued a nationwide non-congregate feeding waiver for 2020-2021 that expressly suspended this requirement. Her testimony characterizing congregate eating as required without noting the waiver is either uninformed or deliberately misleading. The absence of children physically eating at a site during an inspection is entirely consistent with legitimate operation under the non-congregate waiver. (2) UNITIZED MEALS: Honer testified meals must be unitized (all five components, proper serving size, ready to eat). USDA issued meal pattern flexibility waivers during the pandemic specifically allowing substitutions and modified meal components. The defense described bulk dry goods packaged in multi-day bags — which was permitted under USDA's meal pattern flexibility waiver. (3) NO-PROFIT RULE: Honer's blanket assertion that no profit can be made from these programs directly conflicts with the program design that intentionally enrolled for-profit vendors. The nonprofit food service account requirement applies to sponsors, not to for-profit vendors receiving payment for food delivery. (4) ATTENDANCE CONFLATION: Honer described attendance (individual child names/enrollment) as required for CACFP At-Risk programs. She had not yet reached SFSP open sites, but if she applies the same attendance/roster framework to SFSP open sites, that is flatly incorrect — open SFSP sites require only a meal count, not individual identification. (5) RULE 701 VULNERABILITY: Every regulatory characterization is subject to challenge as expert opinion testimony not properly disclosed under Rule 702. Ian Birrell preserved this objection before Honer took the stand. (5) AT-RISK ENROLLMENT — REGULATORY CORRECTION: Honer testified at p. [see Vol V testimony] that CACFP At-Risk afterschool required individual child enrollment and attendance records ("names of the children who are enrolled in that after-school program"). THIS IS FACTUALLY INCORRECT AS A MATTER OF FEDERAL REGULATION. CACFP At-risk afterschool is an open-site model — it serves any eligible child who appears at the site in an eligible area. No individual enrollment roster is required. This is structurally parallel to SFSP open sites. If Honer characterized At-risk as requiring individual enrollment records, that constitutes false testimony about the regulatory requirements. The prior defense analysis flagged this as a conflation issue but the correct framing is stronger: there was NO enrollment requirement for At-risk sites, period. Defense counsel must develop this with a CACFP regulatory expert.

Vulnerabilities Knez's observations are entirely circumstantial. She has no knowledge of what ThinkTechAct/Mind Foundry's actual business operations were, whether they used cloud-based systems (no need for physical files), or whether a consulting/program-administration business would normally maintain heavy paper records. The 'no evidence of a $20 million business' opinion is a Rule 701/702 lay-opinion characterization that was objected to (overruled) but is vulnerable on appeal.

Vulnerabilities The passport fraud subplot is largely resolved — Ismail already pled guilty. The government is using it to paint consciousness-of-guilt for the broader food fraud case. But the relevance is limited: (1) Farah never fled; (2) Ismail's destination (Kenya, where his family lived) and the timing (Eid) offer an innocent explanation. Hostetter did not review Ismail's travel history, did not inventory the luggage, and confirmed the FBI had notice of Ismail's family in Kenya. The Farah passport application is separate and weaker — Farah never fled, and the 'lost at unknown location' claim is legally insufficient to prove guilt in the food fraud case.

Vulnerabilities Kary's cross-examination opportunities (not yet exploited in this volume) will be significant. Key vulnerabilities to develop: (1) He admitted he did not do a full financial analysis himself — the forensic accountants did. He is testifying to conclusions drawn from documents rather than from independent financial analysis. (2) His characterization of the email waiver forwarding (G-48, Farah forwarding MDE bulletin to Ibrahim) as evidence of scheme awareness is highly ambiguous — staying informed about program rules is not evidence of fraud. (3) He did not admit whether FBI interviewed Kara Lomen — the sponsor executive who controlled submissions and who has not been charged. This is a major gap. (4) The Afrique pitch deck shows plans from December 2020 — before Kary claims to have identified fraud; the government must show that the food service revenue contemplated was in fact fraudulently obtained, not that planning to earn food service revenue is itself criminal. (5) The consistent round-number meal counts (2,000/day, 1,000/day) that the government treats as suspicious could also reflect sponsor-level administrative caps set by Partners in Nutrition, not fabrication by site operators.

Vulnerabilities Roase presents several significant vulnerabilities that defense must exploit on cross. First, her food expense analysis ($3.1 million for Empire Cuisine) was self-described as 'conservative' — she credited Lincoln Trading halal meats and Capital Imports coffee/tea/syrups as 'food' even after testifying they did not match children's program menus. This inflates the appearance of fraud by crediting non-program food. Defense should force her to distinguish between restaurant food purchases (legitimate business expense) and program food purchases. Second, Roase has no firsthand knowledge of whether meals were actually served — she is a financial analyst, not a program monitor. She cannot testify to whether children received food; she can only testify to bank records. Third, her characterization of all food program money as 'Federal Child Nutrition Program money' throughout was objected to under Rules 701/702 (foundation/expert) by Mr. Mohring; the Court ruled it went to cross rather than foundation, meaning defense has latitude to challenge how she determined which funds qualified. Fourth, Roase described the USDA pandemic waiver period (COVID 2020-2021) as a period of 'concern' due to increased sites and claims, but she offered no testimony about whether the sites' operations were consistent with pandemic waiver rules (non-congregate, parent pickup, etc.) — her analysis simply assumes the claimed meal counts were fraudulent without addressing whether they could have been legitimate under COVID waivers. Fifth, the 'dormancy then activity' narrative she used for ThinkTechAct, MIB Holdings, and Bushra does not account for the fact that COVID-19 dramatically changed the program landscape in April 2020 — entities that had no prior food program activity would naturally show this pattern upon first enrollment. Sixth, Roase's conclusion that money 'did not go to food' never accounts for cash purchases — she only examined bank account withdrawals and checks, not cash-based food procurement. Seventh, she acknowledged Lincoln Trading purchases (halal meats matching the Empire Cuisine restaurant menu) were credited as food expense even though she believed they were for the restaurant, not the program — but she still counted them in the program-funded food expense total, muddying her own analysis.

Vulnerabilities Frank is easily the most vulnerable government witness type: a borrowed agent with no case knowledge. His admissions were numerous: (1) he had no involvement in the investigation other than a 1-2 week pre-search engagement; (2) he read the warrant affidavit once and cannot vouch for any of the underlying facts in it; (3) he has no personal knowledge of meal service, program rules, or who actually controlled what entities; (4) he could not explain the $3.52/$2.01 reimbursement rates or how they were set; (5) he made the search warrant affidavit's government theory ('almost none of this money was used to feed children') sound like an established fact on direct, but it is merely an allegation from a probable-cause document. Critically, his direct examination was far broader than it needed to be — he was allowed to opine on whose documents were whose, who lived where, and what meal count forms meant — all beyond his actual knowledge. The government's decision to use him to introduce the meal count comparison was a significant miscalculation: an agent who did not 'scrutinize these closely' testifying that forms looked 'identical' is weak authentication that invites reasonable doubt about his analysis. There are also no regulatory characterizations about CACFP/SFSP requirements from Frank — he admitted he has no detailed knowledge of how the program works.

Vulnerabilities Ball is a pure custodian witness with no knowledge of program rules, regulations, or the meaning of any document he handled. He cannot authenticate handwriting, identify document authors, or connect any document to any particular defendant's knowing participation. The duffel bag ownership is unproven. The government used him to flood the jury with documentary evidence through a witness who lacks foundation to contextualize it. No COVID-19 waiver issues arose directly with this witness, but the box trucks cross-examination opened that door. Key vulnerability: Ball confirmed no handwriting analysis and no investigation of Mahad Ibrahim's email address.

Vulnerabilities Pure technical foundation witness; no substantive knowledge of case. The Mac laptop gap is the most significant — its contents were never obtained. Defense should demand a full accounting of what devices were seized versus what was actually imaged and processed. Cloud data gaps mean the Cellebrite reports are incomplete records, not comprehensive captures of all defendants' digital activity.

Vulnerabilities Pitzen was not cross-examined in this volume, so vulnerabilities must be assessed from what will arise. Key vulnerabilities include: (1) He was assigned to the case in early 2023 and was not the follow-the-money agent — the FBI forensic accountants did that work. He is a conduit for phone evidence he didn't create or control; (2) He repeatedly mislabeled Partners in Nutrition as 'Partners in Quality Care' throughout his testimony (both names appear in the record, but the organization operating under MDE as a sponsor is Partners in Nutrition — this confusion about the actual sponsor entity name could affect exhibit foundation arguments); (3) He testified to inferences from text message content ('I found messages coordinating distributing the money, not so much about messaging about coordinating food distribution') that reflect his interpretive gloss, not objective fact — the text messages speak for themselves and show both money discussions AND operational references; (4) The 'no logistics' argument is overstated — the texts do reference menus, menu matching, MDE approval processes, site visits, sponsor check pickups, and food distribution operations; (5) The regulatory framework governing rosters and SFSP open sites needs to be challenged: Pitzen testified that 'initially some of the sites were — during the summer food program, didn't need a specific name. After the stop pay they needed to have rosters' — this characterization should be scrutinized against actual SFSP open-site regulations, where individual-name rosters were never required.

Vulnerabilities Parks has devastating vulnerabilities that were never explored due to the complete absence of cross-examination. The critical failures are: (1) Parks has only 3 months of post-training experience and was explicitly 'still in training status' — he had never led an investigation and was assigned to assist on this case; (2) Parks never testified that he reviewed any USDA regulations, program rules, or the specific agreements governing whether the sites at issue were SFSP open or closed enrolled sites — the entire school-enrollment comparison analysis rests on the unstated assumption that children served must be enrolled in local public schools, which is flatly wrong for SFSP open sites; (3) Parks never disclosed his methodology for name-matching — whether he used exact full-name matches, partial matches, phonetic matches, or some other standard; different children with the same common name (e.g., there are thousands of Muhammads and Fatimas in Minneapolis) could legitimately produce matches or mismatches; (4) Parks never addressed the COVID pandemic waivers, under which sites could distribute packaged meals for pickup by parents and guardians without any on-site attendance tracking — meaning rosters could legitimately include children who never appeared at a specific meal site but were enrolled in the program; (5) the school district records Parks obtained covered only 20 of the roughly 330+ Minnesota public school districts, and covered only public school students — children in charter schools, private schools (including Islamic schools, which would be disproportionately represented in the Muslim community at issue), homeschool programs, pre-K programs, and recently arrived refugees not yet enrolled would not appear in Parks's database of 193,249 names; (6) Parks never testified that he reviewed whether any of the named children in the rosters were real individuals outside the public school system; (7) the 'peculiar names' evidence (Serious Problem, etc.) while compelling, was never cross-examined on chain of custody or the possibility that these were IT system placeholder entries for incomplete records rather than deliberate fabrications; (8) Parks was relying on Kara Lomen's operational guidance (e.g., G-241, G-136) as his understanding of what the program required — but Lomen's guidance reflected the sponsor's internal policies, not necessarily actual USDA regulations, and the two could differ; (9) Parks never testified about the actual meal claims submitted — he analyzed only rosters, not whether meals were actually purchased and distributed; there is a gap between a flawed roster and a fraudulent claim for meals never served; (10) in Trial 2, Parks admitted under cross that he did not review USDA regulations and relied solely on MDE witness representations.

Vulnerabilities Blackwell is the government's most polished financial witness but has several significant vulnerabilities Defense counsel should exploit on cross. (1) The FIFO methodology she admitted to on voir dire requires her to 'show all sources' when funds are commingled — meaning she cannot actually prove that food program dollars specifically purchased any asset, only that such dollars were in the account. Defense should demand, on cross, her account-by-account analysis: what other sources were in the Empire accounts? What non-program revenue did Empire or ThinkTechAct receive? (2) She admitted her methodology is based on specialized training — undercutting the government's claim she's just 'doing math.' (3) The demonstrative charts (N-series) will not be in the jury room, meaning jurors must reconstruct the money-flow arguments from memory. (4) She has not reviewed whether any of the real estate purchases involved legitimate business investment intent or generated revenue. (5) The $8.2M figure is a subset of $41M — what happened to the other $32.8M? Defense can argue the program was mostly legitimate and the fraud fraction, if any, is narrow. (6) Partners in Nutrition (Kara Lomen's organization) is identified throughout as a primary payment source — yet she was never charged, never interviewed, and her internal decisions about approving claims are not accounted for. This is a major gap in the causation chain.

Vulnerabilities Blackwell is methodologically vulnerable on multiple fronts. Her FIFO tracing methodology is an unchallengeable choice only if not contested by a competing expert — using LIFO the entire analysis could yield different attribution. Her demonstrative charts (N-135a, N-135b) are deliberately selective: they show only federal program money flowing in, omitting the $300,000+ landlord improvement allowance, the ~$600,000 net investment income, and the large reverse flow of $447,102 out of Afrique back to Empire Enterprises nine days after the government's traced deposit. Her background (law degree in Romania, Romanian National Police, then FBI forensic accountant) is solid on credentials but she has never worked in the private sector or participated in a commercial build-out project, making her susceptible to cross on commercial real estate funding structures. She made a self-corrected error on direct (calling Abdimajid Nur an owner of Empire Cuisine), which is an impeachment touchstone for closing. She conceded she does not know who actually initiated the wire transfers — only that the entity sent them and Nur was a signatory. She is not an expert witness (she confirmed she is a fact witness), limiting the opinions she can offer. No regulatory characterizations appeared in this portion of the transcript.

Vulnerabilities Ahmed is a deeply compromised cooperating witness with substantial credibility problems. He personally committed over $2 million in fraud using two fraudulent operations (Southwest Metro Youth and Free Minded Institute under PIN), accepted $1M+ in kickbacks, created and signed multiple false corporate documents, fabricated invoices, and backdated contracts to obstruct a federal investigation — all before cooperating. He only came forward after being contacted by the FBI, and is testifying to reduce a 4-5 year sentence. His testimony contains internal inconsistencies: he claims to have emailed Emily Honer at MDE about fraud in December 2020 (a highly significant claim that was not developed by either side), yet continued committing fraud himself for months afterward. His knowledge of individual defendants' intent is largely inferential — he received kickbacks and infers they paid them for fraudulent protection, but he has limited direct evidence of what Abdiaziz Farah, Said Farah, or Abdimajid Nur actually knew about the underlying meal count fraud at their sites. His characterizations of Partners in Nutrition/Kara Lomen are limited to knowing her as a sponsor executive and noting that Julius Scarver operated as her right-hand man in the same way Ahmed operated for Bock — this does not establish PIN fraud. Ahmed consistently mislabels the organization as 'Partners in Quality' or 'Partners in Quality Care' — if this testimony is used to imply PIN involvement in fraud, the name confusion could be exploited. His Free Minded Institute work under PIN involved smaller-scale fraud with different mechanics and limited self-knowledge ('I will say maybe — if we claim 200, probably some were fake').

Vulnerabilities The most serious vulnerability is that Vaaler had not reviewed any case evidence. This is a fundamental problem for an expert witness — the jury knew he had no connection to the actual facts. On redirect, Mohring rehabilitated him somewhat by establishing that transactional informality extends beyond money transfers to all business dealings. However, the government's core attack — that his remittance/microbusiness framework describes $200/month transfers, not millions — was largely unanswered. For defense counsel's use: if deploying a similar expert, the expert must be prepared to engage with the actual transaction scale and explain why the cultural framework still applies at larger dollar amounts. A better-prepared expert would have reviewed at least the Afrique financial records and offered an opinion connecting the cultural practices to specific facts in evidence.

Vulnerabilities Menozzi is being used for implicit comparison purposes, but his testimony cuts both ways. His Shakopee district has a large Somali population — actually supporting the plausibility of community demand. His daily meal figures (300-600 at the high school) reflect an institutional model with cafeterias, cooking staff, and students eating on-site — fundamentally different from SFSP non-congregate open-site parental pickup under COVID waivers. Defense must hammer the distinction: (1) Menozzi's model is NSLP/school meals, not SFSP; (2) Menozzi's students ate on-site in cafeterias, while SFSP open sites distributed packaged meals for parents to take home for all their children; (3) Menozzi's district has a fixed enrolled population — SFSP open sites have no enrollment limit and can serve any child in a low-income area. Menozzi's testimony about Shakopee's COVID approach (pickup at school plus bus delivery to low-income areas) is actually consistent with the types of non-congregate distribution the defendants engaged in. The $4.5M/7800-student math also works against the government: that is $577 per student per year, or roughly $3.15 per student per meal day — consistent with per-meal reimbursement rates, not evidence that large-scale meal service is implausible.

Vulnerabilities Van Horn's testimony was narrow — she only established the interstate nexus via Arkansas servers. She had no knowledge of the underlying transactions or whether the payees were legitimate. The construction/remodel check to Johnson-Reiland and the Candy Cove Trail real estate purchase could have legitimate explanations that were not explored. The defense failed to probe whether these were proceeds of fraud or legitimate business expenditures.

Vulnerabilities Graffunder is the government's most complex witness because she actually confirmed food distribution was happening at the site — she just said it wasn't 'a lot' of people and the food was vegetables in boxes, not prepared meals. Her testimony actually corroborates that the Acacia Montessori / Scott Park area was a food distribution point. The government used her to attack the volume of claims, but she unwittingly confirmed the program was real. She did not count people, did not watch continuously, and her cameras had coverage gaps. The FBI's failure to collect her 2021 surveillance footage is a serious investigation flaw. Her characterization of the food as 'vegetables and tomatoes' rather than prepared meals may actually support the defense — CACFP meal distribution during the pandemic included meal kits and multi-day packages, not necessarily hot ready-to-eat food. Her testimony that 'it was not every Friday' and she saw this 'four or five times' reflects her memory from three years earlier, asked for the first time in March 2024.

Vulnerabilities Wade-Ardley is an emotionally sympathetic witness — genuinely well-meaning, clearly duped. Her vulnerability for the defense is significant: she has zero knowledge of the Dar Al-Farooq site, which was the primary high-volume site in the claims she was shown. She confirmed she would distribute food to anyone who came up, without checking IDs or verifying child counts — behavior consistent with open-site SFSP rules. Her emotional distress at 'being taken advantage of' and her admission that she 'doesn't like lying' are powerful for the government, but defense counsel effectively neutralized the 3,000-number issue by establishing she accepted Shariff's explanation at the time and had no reason to doubt it. The government's use of her emotional reaction to checks she knew nothing about (Empire/Wadani checks) over repeated defense objections was arguably 403 territory — this line should be preserved on appeal. She had no knowledge of the regulatory framework (waivers, SFSP open-site rules) and freely admitted it on cross — undermining any government inference she draws from the mere fact that large numbers 'seem high' to her.

Vulnerabilities Wanless is the most vulnerable government witness on this day from a regulatory standpoint. She has zero knowledge of CACFP/SFSP rules, COVID-19 waivers, or non-congregate service. Her negative observations are not probative of fraud if meals were distributed via non-congregate pickup (which pandemic waivers explicitly permitted). She only worked through mid-July 2021, leaving the bulk of the claimed fraud period outside her observation. She was contacted mid-trial, read about the case before testifying, and cannot exclude authorization through her corporate superiors. Ian Birrell also correctly noted she does not know who submitted the enrollment forms or what a sponsor does versus a vendor.

Vulnerabilities Same structural weakness as Wanless: zero knowledge of CACFP/SFSP rules, COVID waivers, or non-congregate service. Theisen also had limited hours (7am-6:30pm, Monday-Friday primarily) and could not observe weekend or off-hours distributions. The mathematical cross by Carlson effectively neutralized his 'I never saw it' testimony. He also confirmed real food distributions did occur at La Cruz via the unidentified box van. The government over-relied on a building maintenance worker to prove program-level fraud.

Vulnerabilities Same limitations as Walker: weekday-only, inside park-perimeter only. Ruhland's physical-evidence argument (tire marks, trash, noise) assumes the claimed meals would all be distributed from inside the park — but if they were distributed from the parking lot of a nearby business or roadside, these physical markers would not apply. His testimony about gravel parking lot degradation is based on experience with large events like 3,000-person food truck festivals — a sudden daily surge of 200 vehicles would indeed leave marks, but his observations during COVID may be compromised by the already-increased foot traffic he described. Additionally, Ruhland split time with Hyland Park starting around Thanksgiving — he was personally present at The Landing only about three days per week in December 2020, when the highest claimed counts (14,260 meals/day) were occurring.

Vulnerabilities Mills was a narrow records custodian; he had no knowledge of the purpose of the $250,000 payment to Ikram Y. Mohamed or whether it related to food program fraud. Who is Ikram Y. Mohamed and what was the payment for? That line was not explored.

Vulnerabilities Morris was a limited-scope custodian who only reviewed the specific charged emails — he never read the search warrant, did not personally pull the data, and could not speak to Google's full production. He admitted the exhibits did not include IP addresses (which would identify actual sending locations), metadata about folder placement, or read receipts. The government structured his testimony narrowly to avoid these vulnerabilities. A future defense team should consider obtaining the full Google production (including IP logs and metadata) through discovery, which may show where emails were actually sent from or received.

Vulnerabilities Rieta is a technical custodian with no knowledge of the broader fraud scheme. She cannot speak to what the proceeds of the wire transfers represent or whether the underlying transactions were legitimate. The defense did not challenge whether the purpose-of-payment information (real estate) on the Farah wire was supplied by Farah himself or by the title company. No foundation objection was raised to the branch identification of Farah as the walk-in customer — her identification was based on reviewing the account records, not independent verification.

Vulnerabilities Lopez is the weakest of the government's site witnesses because he inadvertently confirmed that food distribution was real and ongoing at the location, that Somali Community Resettlement was actively packaging food there, and that significant food waste appeared in the dumpster. His testimony was primarily that he did not see large crowds — but he was not present on weekends, not present before 8 a.m. or after 4 p.m., and did not count or track the people he saw. He also admitted he saw multiple families per week picking up bags. His math-based concessions to Carlson suggest that with multi-day meal bundles for multiple children, his observed traffic of 5-10 people per week is not grossly inconsistent with the claimed totals. The 3-year delay in the FBI contacting him is also notable.

Vulnerabilities Amenta is a technical custodian with no knowledge of or opinion about the underlying transactions. His only function is to establish the interstate element and authenticate records. The defense could challenge whether the government's exhibit L-43 was produced just before trial (certified April 22, 2024, weeks before this trial day), suggesting last-minute subpoena compliance. No foundation attack was made on the underlying transaction records themselves (chain of custody from the originating banks). The defense could also explore whether CHIPS rather than Fedwire was used for any of the international transfers, which might have different interstate nexus characteristics.

Vulnerabilities Walker's testimony is limited in scope: he was present Monday-Friday during park hours, not on weekends, not before 8:30 a.m., not after 5:30 p.m. The claim application listed meals served 11:00-12:00 — a window that overlaps with Walker's schedule, but he admits he would not know what happened outside the park perimeter. The government pushed him to say 'it would be difficult to imagine' he wouldn't notice large groups — but his actual testimony is 'I believe I would have' and 'I don't think so,' which is softer than the government's narrative. The 3-year delay in the FBI contacting him is a significant vulnerability — either the FBI knew the site was fraudulent from other evidence and did not prioritize confirming it, or they were building the case without corroborating basic facts. Walker has no knowledge of what happened outside the park. He also cannot explain the mechanism of how claims were submitted or who submitted them — his testimony is purely about absence of observation.

Vulnerabilities Shariff's decision to testify is the highest-risk strategic move in the case. The government's cross will focus on: (1) his concealment of the food program from attorney Jacob Steen; (2) the pitch book/investor materials showing the food program as a revenue stream for Afrique; (3) the $900,000+ flowing through his entities; (4) the Kara Lomen 'turn off the errors' text message and Shariff's relationship with PIN; (5) his relationship with Mahad Ibrahim and the January 2021 meal count submission before Sysco deliveries began. Watch Vol XXVI closely.

Vulnerabilities Haji's Saturday-only testimony actually helps the defense on the 'food was real' point but inadvertently helps the government's fraud theory by establishing that distribution only occurred one day per week, not daily as the program claimed. For Defense Counsel: this witness is a double-edged sword. He proves real food was distributed but also supports the government's argument that the program claimed reimbursement for days on which no food was served.

Vulnerabilities The friendship bias is real and will be argued by the government. More problematic is the inconsistency with Adan and Haji: Kariye says food was distributed many days of the week, while Adan and Haji (who were also there consistently) say they only saw it on Saturdays. This internal defense inconsistency undermines the credibility of the broader defense narrative. For Defense Counsel: either reconcile this inconsistency before trial or carefully manage whether to call multiple eyewitnesses whose accounts may diverge on the frequency of distribution.

Vulnerabilities Adan is a genuinely helpful witness as a food recipient — the government cannot easily call her a liar. However, her testimony is limited to Saturdays, which creates an inconsistency with the daily meal count claims for the site. For Defense Counsel: use Adan for the 'food was real' purpose, but address the daily vs. Saturday distribution issue proactively.

Vulnerabilities Petracek is clearly not an expert in CACFP/SFSP and his 'absurd numbers' testimony is based purely on his lay view of Lexington's size — he has no knowledge that SFSP open sites may draw from broad geographic areas under program rules, or that COVID waivers allowed non-congregate pickup potentially serving children traveling from Blaine (70,000 residents), Circle Pines, or Mounds View. He confirmed MDE never contacted the city, which is a significant regulatory failure that cuts both ways. The fact that the application used the mall address rather than 'Tot Park' address also creates some factual complexity the government glossed over.

Vulnerabilities Hamilton was only present weekday business hours (10 a.m. to 4 p.m.) and was gone evenings and weekends. The defense has photographs showing what appears to be food distribution from a box truck in the alley adjacent to the Cedar plaza. The Cedar Square Towers directly adjacent to the site house thousands of low-income Somali families with children — the program's precise target population. Hamilton's own cross-examination established demographic facts that support the plausibility of the program. He acknowledged a prior, legitimate relationship with Mind Foundry. His 'I never saw it' testimony is limited by his schedule and vantage point.

Vulnerabilities Czapiewski became a vehicle for the government's hub-and-spoke fraud theory. The January 2021 discrepancy is a serious vulnerability that prior defense counsel did not neutralize — they needed to either explain what food was used for January or attack the government's meal count records. The multiple checks from other entities were devastating; defense redirect did not effectively address them, only noting Czapiewski did not know what the checks were for. For Defense Counsel: if Czapiewski or similar Sysco witnesses appear, develop a clear line of questioning establishing whether the food volume purchased is consistent with the total claimed serving numbers, and be ready to explain the pre-Sysco period.

Vulnerabilities Nelson admitted authorization was given by her manager, saw one actual distribution (truck), and has no CACFP/SFSP knowledge. Her testimony does not rule out additional distributions on days she was not present (part-time, flexible schedule). The surveillance footage gap is a major problem for the government. The government never called Kelly Krick (the manager who authorized the program) — a significant gap.

Vulnerabilities Hawk only worked mornings (7:30-noon), was not present on weekends, and has no knowledge of CACFP or SFSP program rules, COVID-19 waivers, or how meals are defined under the program. Her 'I never saw it' testimony does not foreclose the possibility of weekend or evening distributions, non-congregate pickup, or distributions when she was not present. She is personally a low-income resident who admitted she would have gotten free food if it were available — making her a sympathetic witness for the government but also one whose knowledge is severely limited by her schedule and position.

Vulnerabilities Steen was a well-intentioned defense witness whose testimony backfired. The critical error is that defense called a witness who, under cross, became the government's most effective vehicle for arguing consciousness of guilt. The attorney-client privilege ruling (defunct company loses privilege) is a legal issue that prior counsel either did not anticipate or could not avoid. The better strategy would have been to brief the privilege issue more aggressively before trial, or not call Steen at all and rely on documentary evidence of Afrique's legitimacy without opening the 'did Shariff disclose the food program' door. The testimony also inadvertently highlighted the $900,000 funding gap at the Afrique project.

Vulnerabilities Garcia gave prior inconsistent statements to the FBI (someone did ask permission) versus her trial testimony (no one asked). The signed D4-1 form by her own assistant grants ThinkTechAct Foundation permission to operate CACFP programming at the site on August 30, 2021. She has no CACFP/SFSP knowledge. She saw box trucks on multiple occasions at the site. She pulled down a sign advertising the food program, which proves someone was affirmatively advertising the program at the site. The FBI never retrieved surveillance footage that could have resolved the question definitively.

Vulnerabilities Omar's story about who asked him to draft the email changed during cross-examination, which the government correctly exploited as a credibility issue. The Wade-Ardley reply email ('maybe 1,000 each week') is very damaging because it comes from a neutral third party contemporaneous with the program. For Defense Counsel: consider whether to call Wade-Ardley as a defense witness to explain the full context of the email exchange, or investigate whether the '1,000 each week' figure was for Oak Grove only (and Dar Al-Farooq was separate). On redirect, defense attempted this contextualization but did not fully close the gap.

Vulnerabilities Hassan's $250,000 receipt from Feeding Our Youth seriously compromises her role as a credible 'legitimate investor' witness. If the jury concludes that payment was a distribution of fraud proceeds rather than a personal loan, her entire narrative collapses. The government did not need to prove this definitively on cross — the association alone was damaging. Additionally, she has a financial interest in the outcome to the extent she wants to recover her investment, which a skilled prosecutor could use to color her testimony as self-interested.

Vulnerabilities Major gap: she was not employed at Heather Court before mid-July 2021, so she cannot testify to anything that happened January through mid-July. She worked limited, flexible hours with no weekend presence most of the time. She has no CACFP/SFSP knowledge and cannot speak to program rules or COVID waivers. The government presented her as if she covered the full period, which is misleading. No surveillance cameras means no corroborating or contradicting footage.

Vulnerabilities Ali is a sympathetic witness — a genuine investor who believed in the project — but he is effectively a community character witness for the Afrique concept, not for Shariff specifically. The government's $900,000 question was sharp and unanswered. Ali also created a small inconsistency problem on the name-change testimony.