Architect of the iFeed methodology · founder of iFeed · production floor since 2017
I sit at the precise intersection that regulated life sciences now needs and almost no one occupies: production-floor work across CRO, sponsor, and MedTech environments since 2017 — bioanalytical method validation, bioequivalence study oversight, clinical trial operations, quality systems, and most recently medical-device quality — threaded through with intentional intervals spent stepping back from the floor to strengthen the foundation, master the AI and tooling layer, and synthesise what the floor was teaching into a methodology that could be written down and run as code. The methodology you read on this site is what those layered years assembled.
iFeed is the practice. The library is the substrate. This page is mine, and I am putting my name and my work on it.
The regulated industries already have the vocabulary for absorbing agents that bring possibility and vulnerability in the same step. That vocabulary lives in pharmacovigilance, in quality management systems, in the ICH guideline lifecycle of every medicine ever approved. AI inherits that vocabulary — modified for non-deterministic outputs, modified for the regulatory regimes (EU AI Act, FDA's evolving SaMD posture, MHRA's AI airlock) that have just finalised or are about to.
The combination that 2026 needs is operational regulated-substrate at depth, plus AI-native execution at the frontier, plus a documented methodology that bridges the two without losing either side. That combination is rare. I have built mine in the open across pharmacy education (2011–17), production-floor years from 2017 forward, and intentional intervals spent synthesising what the floor was teaching — and I am claiming it here without apology.
Bioanalytical method validation through ICH M10. Bioequivalence study oversight across multi-jurisdictional pathways (FDA, EMA, CDSCO, ANVISA). Clinical trial operations across Phase I–IV. Quality governance across GxP frameworks, 21 CFR Part 11, ALCOA+, GAMP 5, EU MDR / ISO 13485, and now EU AI Act Annex IV. Built from 2017 forward across CRO, sponsor, and MedTech environments. The substrate is documented — not claimed.
The methodology is the policy layer between regulator-grade rigour and AI-native execution. It is its own intellectual property — PQIOS framework, Antidote+Vaccine governance pattern, the operating loop. Already running; documented openly. The methodology travels with the practice and outlasts any single engagement.
Fluent with agent orchestration, MCP-standard tool ecosystems, multi-agent workflows, local-first model deployment, the patterns that turn methodology into shipping product. The execution layer is no longer the constraint — the substrate-plus-methodology layer is. That is exactly the layer I bring.
Methodology without artefact is rhetoric. The iFeed methodology stands on three concrete builds, each delivered against a real regulated brief, each leaving a documented quality system and an operational team behind it. These are the builds I point to when someone asks what the substrate actually produces.
Architected and stood up the full quality management system for a clinical-research and bioanalytical operation in ninety calendar days — SOP architecture, change control, deviation handling, CAPA, training infrastructure, audit-trail discipline. The system passed sponsor audits and held under inspection pressure. The 90-day artefact is now the reference QMS shape I deploy from.
The clinical-trial assurance platform — site oversight, monitoring quality, ICH E6(R3)-aligned risk-based quality management codified into a working system. Designed to become the operational backbone for sponsors and CROs running multi-site trials with regulator-grade evidence requirements. ClinAssure is being built because the field still treats RBQM as a slogan instead of a system. Currently in active development; not yet available for engagement.
The Quality Intelligence Operating System — the agentic, MCP-native execution surface where the iFeed methodology will live as code. The PQIOS framework, the Antidote+Vaccine governance loop, the audit-grade record of every reasoning step. QI_OS is what will make the methodology travel from page to practice without losing the audit trail regulators will eventually require. Currently in active development.
The three builds are not interchangeable. If you are a CRO or sponsor needing a QMS that holds in inspection — Build 01 is the reference. If you are running multi-site clinical operations and need RBQM that is a system, not a deck — Build 02 is the answer. If you are building AI-touched workflows in a regulated context and need governance that is auditable end-to-end — Build 03 is the operating layer. Read the timeline below to see how the substrate that produced these was assembled.
The substrate did not appear. It was assembled across pharmacy education, production-floor years across CRO, sponsor, and MedTech environments, and intentional intervals where I stepped back from the floor to strengthen the foundation, master the AI and tooling layer, and synthesise what the floor was teaching into something that could travel beyond a single employer. The phases below are the honest record — production years and synthesis intervals are interleaved, not sequential, and the page reads them that way.
The full pharmacy curriculum — pharmacology, pharmaceutics, pharmaceutical chemistry, pharmacognosy, regulatory frameworks at the entry layer. The educational substrate that every production-floor year afterwards runs on. Pharmacy is not the same field as data science with a pharma label; the difference shows up later in how the methodology reads regulator language.
Master of Science in Pharmacy at NIPER — India's premier pharmaceutical research institution. NIPER's research-grade orientation is what later translates into methodology architecture rather than process-compliance reflex. This is also where the habit of reading primary literature, ICH guidelines, and regulator notices as living documents was set.
Entry to the regulated production floor. Bioanalytical method validation and bioequivalence study oversight at production scale — multi-sponsor, multi-jurisdictional dossiers, LC-MS/MS method transfer, ISR (incurred sample reanalysis) discipline, the uncomfortable conversations when a study trends toward an out-of-specification result. The bioanalytical instinct was built here, and the first lesson the floor teaches sticks: in regulated work, every action either generates audit-trail or it didn't happen.
Architected and stood up the full quality management system for a clinical-research and bioanalytical operation in ninety calendar days — SOP framework, change control, deviation and CAPA workflow, training matrix, audit-trail discipline. The system passed sponsor audits and the reference shape for every QMS I deploy from afterwards crystallised here.
Intentional intervals away from the floor — spent strengthening the pharmacy and regulatory foundation, mastering the AI and tooling layer (agent orchestration, MCP-standard ecosystems, local-first model deployment, the frontier patterns that turn methodology into shipping product), and synthesising what the floor was teaching into a written, legible, transferable methodology. The iFeed continuation across 2019–2023 lives in this layer. These are not gaps in the timeline; they are the layer that turns operational practice into intellectual property.
Bioequivalence study governance at scale — protocol oversight, sponsor-CRO quality interface, regulator inspection readiness across FDA, EMA, ANVISA, CDSCO submissions. Then the sponsor side of the table: quality oversight from the company that owns the molecule, sees the dossier, and carries the regulatory consequence. The change in vantage point that completes the bioanalytical–bioequivalence–clinical picture.
The medical-device substrate. ISO 13485, 21 CFR Part 820, EU MDR, Software-as-Medical-Device, the device-software-process triad. Galway as the European MedTech cluster — the geography where the AI-in-regulated-industry thesis acquired its medical-device dimension. The most recent production-floor layer; the one that closes the loop from molecule to device.
The methodology takes shape as code — QI_OS — and acquires its public surface. The substrate assembled across pharmacy education, production-floor years, and synthesis intervals becomes legible as a single operating model. ClinAssure is being built in the same window. The methodology is no longer something carried implicitly; it is documented, and the runnable surfaces are under construction. Both QI_OS and ClinAssure are in active development; not yet available for engagement.
Galway as base — Pune as origin — global as scope. This page, the library, the methodology, the AI section: all of it is iFeed's public surface.
Production-floor depth in clinical-research quality. Bioanalytical method validation. Bioequivalence study governance. ICH-aligned clinical-trial substrate. The discipline of audit-trails and ALCOA+ records. The reality of regulator inspections. I help regulated organisations — pharma, CRO, bioanalytical operations, MedTech — adopt AI without breaking compliance.
Twenty-four services, organised across the quality lifecycle: gap analysis, audits, system design, implementation, remediation, fractional QA, project management, interim leadership. The catalogue is mapped to the four stages of how regulated work actually flows: Assess, Design, Build, Operate.
Gap analysis, audits, vendor qualification, verification. Ten services across the assessment surface.
System setup, QMS architecture, feasibility analysis, continuous improvement. Four services across the design surface.
Implementation, remediation, CSV / 21 CFR Part 11, instrument and equipment QA. Four services across the build surface.
Maintenance, project management, interim QA leadership, embedded support. Six services across the operate surface.
Engagement runs by retainer or scoped contract — structured shape, not hourly billing. The shape of work is the surface I optimise; the commercial conversation happens in private once the fit is real.
The intellectual surface for AI as it touches bioanalytical, bioequivalence, clinical trials, and the medical-device boundary. Three deep dives anchor the section: Governance (the Antidote+Vaccine framework), Agentic AI (the 2027+ thesis), Local AI (the privacy-preserving infrastructure that regulated work actually requires). Audience-routed across academic, research, industry, and business readers — same substrate, different cuts.
What distinguishes the AI section here from the broader commentary surface: every page is read from the regulated operational floor, not from analyst-detached observation. When I write about ICH E6(R3) under AI conditions, I have signed off the SOPs that the AI is supposed to support. When I write about EU AI Act Article 9 risk management, I have walked the inspection that produced the risk register.
The historical pattern around iFeed has been: announce, build silently, return with finished work. That is still the cadence — but I am naming it now, in writing, so it is no longer a vacuum. Public surfaces (this page, the library, methodology, AI section, notes) are where the thinking becomes visible. Between visible drops, the practice is in quiet build phase: signature artefacts, client engagements, methodology refinement.
If a page goes silent for a stretch, that is the cadence working as designed — not the practice abandoning the surface. The next visible drop will have substrate behind it. This is the rhythm I am committing to in the open.
Most of what I think out loud lives on LinkedIn and surfaces here in the library, notes, and AI sections. Subscribe by following; engage by writing back. Or see how the substrate is offered as work.