Cohort learning at the intersection of regulated life sciences and AI quality governance.
iFeed Academy is being built for the practitioners now responsible for AI inside regulated environments — QA leads, regulatory-affairs heads, validation engineers, and AI governance officers across pharma, biotech, and MedTech. Most working professionals trained before this surface existed. The Academy is the structured surface for catching up — and for shaping what the next decade of regulated AI quality actually looks like.
The full curriculum shelf, twelve modules.
5 designed · 1 live · 6 in design · cohort metadata to the rightTwelve modules across the curriculum, named and sequenced. Modules 01–05 are designed and ready. Module 06 · QMSR Transition is the current live-design module the pilot cohort scopes against. Modules 07–12 are in design. Each module ships a real artefact: an SOP, a validation package, a governance dossier — not a certificate.
Why this academy. Why now.
The training gapThe regulatory landscape for AI inside regulated life sciences just shifted — and shifted fast. ICH E6(R3) reached Step 4 on 6 January 2025. The EU AI Act entered into force on 1 August 2024, with high-risk medical-device obligations phasing in through 2026–2027. The U.S. QMSR — FDA's harmonisation of 21 CFR 820 with ISO 13485:2016 — becomes effective on 2 February 2026. ISO/IEC 42001, the world's first AI management-system standard, was published in December 2023. And ICH M10 on bioanalytical method validation has been moving into operational implementation across FDA, EMA, PMDA, ANVISA, and CDSCO from 2024 through 2025.
Five major regulatory instruments — four of them brand new — landed inside an 18-month window. Most QA, regulatory affairs, and validation leaders working today were trained before any of them existed. The training surface they need does not yet sit inside any one university, any one vendor certificate, or any one consultancy's playbook. Practitioners are stitching it together themselves, in production, under inspection pressure.
iFeed Academy fills that gap. It is structured around the actual decisions practitioners now have to make — classify, validate, document, govern, defend — rather than around theoretical frameworks or vendor toolchains. The faculty are working practitioners; the cohorts produce deliverables their organisations can ship; the curriculum tracks the regulator-watch cycle in real time.
This page is the curriculum being scoped now. The pilot cohort and subsequent open-intake cohorts are planned — activation is subject to practitioner-panel interest, faculty readiness, and the substrate being inspection-defensible at the moment of opening. Registering interest does not commit to tuition; it places you on the practitioner-panel list for curriculum review and on the pilot-cohort shortlist when the pilot opens.
The four cohort tracks.
Curriculum · design-stageEach track is built around a coherent decision-set the working practitioner now owns. Tracks run independently and can be taken in any order, though Track 01 is the recommended entry point for participants newly responsible for AI inside a regulated quality system. Cohort size is capped at 8–15 to preserve 1:1 faculty access; deliverables are real artefacts, not exercises.
AI Governance for Regulated Industries.
The end-to-end governance architecture for AI systems inside pharma, biotech, and MedTech — from regulatory classification through management-system certification through inspection-readiness. The recommended entry track.
- EU AI Act high-risk classification · Annex III medical-AI taxonomy
- ISO/IEC 42001 AI management system · certifiable scope design
- FDA AI/ML SaMD · PCCP architecture and the 2024 final guidance
- ICH M10 v2 narrow scope · bioanalytical AI/ML implications
- NIST AI RMF mapping · cross-walking to regulated QMS
- Governance-architecture-by-design · embedding controls upstream
- Incident response, post-market surveillance, model-failure CAPA
- Capstone · full governance dossier for a real or proposed AI system
Validation of Non-Deterministic Systems.
How to design IQ / OQ / PQ for systems that do not return identical output to identical input. GAMP 5 Second Edition aligned. The validation problem AI in regulated industries actually faces — treated rigorously, not avoided.
- IQ / OQ / PQ adapted for stochastic and probabilistic outputs
- Reference test-set design · ground-truth construction and curation
- Statistical envelopes · bounds, tolerances, and acceptance criteria
- Drift detection architectures · data, concept, and prediction drift
- Change-control plans for adaptive AI · PCCP in operational practice
- Capstone · validation package for a candidate AI system
GxP-Quality QMS for AI-Native Operations.
Building — or retrofitting — a QMS that holds up under inspection when the operational substrate is AI-native. ICH Q10 and Q9(R1) integrated with ISO/IEC 42001 and the new QMSR baseline. For QA heads architecting from day one or refactoring from legacy.
- ICH Q10 pharmaceutical quality system · AI-era extensions
- ICH Q9(R1) quality risk management · risk to AI quality
- QMSR transition · 21 CFR 820 + ISO 13485:2016 harmonised
- ALCOA+ for AI training data · provenance, contemporaneity, attribution
- Audit-trail design for inference · what a regulator needs to see
- CAPA loops for model failures · closing the detect-correct cycle
- Supplier qualification for AI vendors · foundation models, MLOps platforms
- Document control, change control, training records · AI scope
- Internal audit programme · AI-specific audit trails and evidence
- Capstone · QMS gap-analysis and remediation roadmap
Cross-Region Regulatory Strategy.
The convergence-and-divergence map. Mapping ICH M10, M13A, and E6(R3) across FDA, EMA, PMDA, ANVISA, WHO PQ, and CDSCO. Submission-package architecture and inspection-readiness across jurisdictions for sponsors running global programmes.
- ICH M10 implementation · FDA, EMA, PMDA, ANVISA, CDSCO comparison
- ICH M13A immediate-release oral solid dosage · bioequivalence harmonisation
- ICH E6(R3) GCP · risk-proportionate clinical-trial conduct
- Submission-package architecture · CTD modules across regions
- Inspection-readiness across jurisdictions · common findings, divergences
- Capstone · cross-region regulatory strategy memo for a real programme
How the cohorts run.
Format · live-async hybridLive cohort calls.
Weekly 90-minute live sessions with the full cohort. Each anchored in a working artefact — a real Annex III classification decision, a real PCCP draft, a real CAPA — not slides. Recordings are retained for participants only.
Bi-weekly 1:1 with Sunita.
30-minute private sessions every two weeks. Working through the participant's own organisational context — their classification call, their validation envelope, their submission package. The 1:1 is where the curriculum becomes operational.
Cohort substrate.
Shared template library: classification matrices, validation-protocol skeletons, audit-trail schemas, governance dossiers redacted from real engagements. The cohort substrate persists post-completion and is updated as regulator guidance evolves.
Structured exercises.
Async exercises between live sessions. Annotated regulator-document reading, redacted case reviews, peer-paired drafts. Participants submit; faculty reviews with line-level annotations before the next live session.
Capstone deliverable.
Each track ends with a real artefact participants ship to their own organisation: a governance dossier, a validation package, a QMS roadmap, a regulatory-strategy memo. Reviewed by faculty; peer-reviewed by the cohort.
Cohort size.
8–15 participants per cohort. Below 8 the discussion thins; above 15 the 1:1 ratio fails. The cap is structural, not commercial — full cohorts close registration and roll the next intake.
Faculty lead.
Practitioner-architectSunita Nawale
BPharm (2011–15) and NIPER M.Sc. Pharm (2015–17), with production-floor years from 2017 onward across CRO, sponsor, and MedTech environments — bioanalytical method validation, clinical bioequivalence, regulated-industry quality systems, medical-device quality — threaded through with intentional intervals spent strengthening the foundation, mastering the AI and tooling layer, and synthesising the substrate into a transferable methodology. The faculty seat is held by the same person who carries the operational substrate.
Sunita's work sits at the practitioner-architect seam: deep enough in operational regulated science to know what an inspection actually asks for, deep enough in AI governance to architect controls that hold up. The Academy curriculum is drawn directly from that working surface — the same substrate that anchors iFeed's methodology, the iFeed library, and the weekly long-form practice on iFeed.com.
Guest faculty for specific modules — senior regulators, audit principals, MLOps leads, in-house QA heads — will be announced ahead of each cohort intake.
Tuition & scholarship.
Pricing · design-stageIndex-priced · scholarship-supported.
Tuition is being indexed to comparable executive-certificate programmes in regulated quality (ASQ, RAPS, ECA Academy, ISPE GAMP 5 training tracks). Specific cohort pricing is confirmed at the point of pilot-cohort intake, not before — the design-stage commitment is to fair indexing, not to a fixed number we may need to revise as scope settles.
- Sliding-scale tuition for academic researchers and full-time students
- LMIC scholarships · partial and full, based on regional purchasing-power adjustment
- Cohort scholarships for early-career practitioners (under 5 years' experience)
- Organisational rates for teams of 3+ from the same sponsor
- Pilot-cohort participants receive lifetime access to cohort substrate updates
Current cohort pricing and scholarship application details are released through the register page at the start of each intake window.
Cohort phases.
Phase-based · subject to interest and feasibilityCohorts move through phases, not calendar dates. Each phase opens when the substrate is ready and the practitioner-panel signal supports activation. We name the phases honestly so registered-interest readers know exactly where the work is — without locking ourselves into a date that becomes a blocker.
Curriculum review.
Practitioner-panel review of all four tracks. Module-level feedback from QA heads, validation engineers, regulatory-affairs leads. Curriculum locks when the panel signal converges — no fixed deadline. This is the phase the page is in today.
Pilot cohort.
Track 01 pilot — up to fifteen invited seats drawn from the practitioner panel and registered-interest list. Pilot tuition is concessional in exchange for cohort-substrate co-development. Activation is subject to interest threshold and substrate readiness; if either is not met, the pilot stays in design until both are.
Open intake.
Track 01 opens for general intake after the pilot closes successfully. Tracks 02 / 03 / 04 phase in afterwards in cadence with practitioner-panel readiness reviews. Each track opens when its module set is inspection-defensible — not on a calendar.
Be among the first fifteen.
Register interest now to join the practitioner-panel review and the pilot-cohort shortlist. No tuition commitment at registration — the panel shapes the curriculum, the pilot tests it. Activation timing is subject to interest and feasibility; we will not announce a date until the substrate is ready to defend it.
What the academy is not.
Honest disclaimersClarity matters more than scope-stretching. The Academy is opinionated about what it covers; it is equally clear about what it does not.
A vendor certificate.
iFeed Academy is not affiliated with any AI platform, MLOps vendor, or regulated-industry software supplier. The curriculum is vendor-neutral by design.
A regulator-issued credential.
Completion is recognised by the iFeed cohort and faculty; it is not a regulatory licence, registration, or government-issued credential of any kind.
A substitute for a QMS.
The QMS track teaches how to design and operate one; it does not provide one. Participants leave with a roadmap, not an installed system.
Legal or regulatory advice.
Curriculum content reflects published regulator guidance and operational practitioner experience. It is educational; it is not a substitute for qualified counsel on a specific submission.
An entry-level programme.
The Academy assumes working familiarity with regulated quality systems or AI/ML engineering. Participants are practitioners; the curriculum starts above first principles.