15+ years building and shipping production AI at the intersection of pharma and enterprise technology. From real-world patient analytics in India to leading cross-functional AI initiatives at Amgen, USA — my career is defined by one throughline: turning rigorous science into systems that scale.
THE journey
1
Optum, India
2010–2016 Real World Data Analytics
2
MS Data Science
2016–2018 Statistical, ML & AI Depth
3
ZS Associates
2018–2022 Pharma AI Consulting
4
Amgen
2020–Present Enterprise AI Leader
WHAT I BRING
Leader
R&D Director leading cross-functional AI across clinical development, biostatistics, clinical ops & IT at Amgen
Built AI team — recruited, developed, and retained talent while shipping production platforms simultaneously
Pioneered self-funding flywheel: consulted internal groups like clients, reinvested revenue to grow — a partner-level operating model inside pharma
Earned alignment across groups that rarely co-invest
Consulting DNA
A trained consultant— structured thinking, client management, and the ability to walk into a room where nobody asked for AI and walk out with a funded engagement.
I don't show stakeholders what a model does. I show them what it means for their P&L, their timeline, or their risk profile. Technical translation isn't a skill I learned — it's the reason my platforms get adopted, not just admired.
Equally credible in a data science architecture review and a VP strategy offsite. That range is what separates platform builders from tool builders.
Builder & Learner
Build-first mindset — every platform started as a prototype, with me being involved hands-on before scaling with a team
Technical range: LLMs, agentic AI, causal inference, Bayesian methods — learned by shipping, not studying
Permanent student by design. New framework drops? I'm testing it against a real use case within a week. Not to chase trends — to maintain the technical credibility that lets me sit at both the engineering table and the leadership table.
Portfolio Overview
My Portfolio — At a Glance
"VP/SVP-level relationships across 6 pharma domains — from clinical development to commercial to enterprise infrastructure — built through delivered value, not org chart proximity"
$5-7mn
Total Portfolio Managed Annually
6+
Active Agentic AI Programs
3+
AI Innovative Platforms
Impact and Influence - Major Focus Areas
Trial Recruitment Universe
ATOMIC, GEO, Photon
Amgen Global Development Operations
~2.5x Enrollment
IMPACT
Redefined patient enrollment & retention process across Amgen's clinical portfolio
Accelerated enrollment velocity ~2.5× through Atomic AI and patient intelligence
Onboarded 80%+ of active studies onto the platform
Rescued 6+ major studies from enrollment delays — directly protecting trial timelines and budgets
Growth & Influence
Built executive sponsorship — persuaded senior leadership to embed data & AI into study operations as standard practice
Secured ~$2M annual innovation funding (YoY) for platform expansion and adoption
Managed cross-TA stakeholder relationships across Oncology, Rare Disease, and General Medicine — proved real-time platform value to each TA head
Pioneered product build culture — built platforms that serve multiple groups with rapid, tailored iteration vs. slow vendor cycles
240+ global users across functions, fielding 1,500+ queries per quarter
4× user growth YoY — organic adoption driven by demonstrated value
Accelerated evidence generation for HEOR, observational studies, and regulatory submissions — faster, more exhaustive, and group-specific
Growth & Influence
Penetrated cross-functionally: Regulatory, Observational Studies, Clinical Development, Clinical Research Medical Directors — each group a separate beachhead
Co-designed features with senior leadership from each function — built exactly what they needed, not a one-size-fits-all tool
Pioneered distributed funding model — secured ~$500K from each group independently, de-risking budget dependency on any single sponsor
Authored enterprise-wide technical evaluation of PDF parsing solutions (LlamaIndex, Redact, MinerU) — adopted as the reference benchmark across teams
Pioneered agentic AI platform adoption — first group to demonstrate production value, catalyzing cross-functional uptake
Served as lead technical advisor for enterprise AI platform pilots and Negotiation — coding tools, agentic workflows, evaluation frameworks
Established team as the go-to AI capability group, accelerating enterprise experimentation velocity
GROWTH & INFLUENCE
Trusted enterprise partner for AI evaluation and POC experimentation — consistently selected for early-stage pilots
Secured seat at key architecture decision forums, representing R&D perspective to shape solutions that serve both enterprise scale and group-specific needs
Oncology, GenMed, Rare Disease AI based workflow for research
Deep Dive - Trial Recruitment Suite (~$3M)
Atomic · Geo · Photon | Global Development Operations · Clinical Development
Atomic — AI Site Selection
ML site scoring across enrollment history, PI profiles, competitive trial overlap, and disease similarity. Gradient-boosted + graph network ensemble. Standard workflow across oncology and general medicine.
Geo — Country Allocation
Optimization engine for site and country allocation — maximizes enrollment speed at target cost and risk tolerance.
Photon — Patient Forecasting & Drop outs
Measuring patient enrollment projects and drop outs in real time suggesting need of more sites to avoid delay and stopping early in some cases to dealy over enrollment
FDA Panel —AI advisory panel on model evaluation, bias, and governance and AI usage in Pharma
Relationship Depth
Executive Sponsor: SVP Global Development, VP Clincal Dev Duration: 3+ years, scope expanding
Platform Flywheel
Deep Dive - Design Intelligence (~$2M)
Amgen R&D Platform · 360° Evidence Generation · US · EU · JAPAC
Evidence of Business Impact
More than 240 Users use the platform, Answered more than 1500+ Queries per quarter
Faster, exhaustive and reproducible Evidence Generation
Answer Business questions from Regulatory, observational, and clinical dev perspective all from the same platform
One of its Kind Protocol Database, for the first time offering protocol guidence to everyone on how to minimize ammendments in particualr TA
Expansion Trajectory
Relationship Depth
Executive Sponsor: VP Observational Studies, AVP Regulatory, SVP Clin Dev Working Partners: Clinical Development · Regulatory · Medical Affairs, HEOR, Value and Access Duration: 3+ years, renewed each budget cycle
Client Groups & Investment
Clinical Dev / Biostats
AI-powered 360° research platform integrating ClinicalTrials.gov, PubMed, NEJM/Lancet literature, and Amgen internal trial records. Protocol design benchmarking, endpoint intelligence, and medical Q&A.
Observational Research
Meta Analysis across various studies, , evidence synthesis for many observational outcome needs
HEOR
Payer evidence generation, comparative effectiveness, value dossier support for market access.
Dynamic physician targeting platform — ML models updating rep call lists weekly across national sales forces. First of its kind of project that scaled to transform how indsutry operates SalesForce
NBA Platform
Next Best Action engines for HCP engagement — cross-channel orchestration across brands, segments, and touchpoints. Full-cycle programs from data strategy through production.
Relationship Depth
01
Relationship Level: Commercial Analytics Senior Leaders, Data Science ED/Dir
02
QBR Presence: Presented to ED-level quarterly business reviews
03
Tenure: 4 years — March 2018 to February 2022
04
Nature: Building AI at Scale, ML roadmaps, Penetrate AI culture via redesigning workflows
Evidence of Business Impact
92%
Recall
vs. prior rule-based targeting models with 40% less saleforce
12%
NBRX increase for Cardio and Inflam indications
Repeat Business & Account Growth
Roche/Genentech: Supported Site Selection and Model buidling for effective enrollment
Cross Sold Medical affiars, Build state of the art NLP based engines for identifying contradiction in market for MSLs
New workstreams (omnichannel, KOL mapping) sold on delivery trust
WORK SAMPLES
01 — Trial Recruitment Universe
AI-driven site selection, geographic optimization & patient retention
02 — Design Intelligence
Research platform — evidence generation, competitive landscape & protocol optimization
Confidential — Prepared for Interview Discussion
Work Sample 01
Trial Recruitment Universe
ATOMIC · GEO · PHOTON
AI-Driven Site Selection, Geographic Optimization & Patient Retention
Work Sample 01 | Strategic Problem & Thesis
The Strategic Problem
Why Clinical Trial Enrollment is a $Billion Bottleneck
80%+
Of trials miss enrollment timelines
Adding 6–12 months to development programs
$5-6M
Per day in delayed revenue
Per late trial, depending on the therapeutic asset
My Thesis
I identified trial enrollment as a systemic problem requiring a connected intelligence pipeline — not point solutions.
I conceived, funded, and led the development of three integrated AI products that together form the industry's first closed-loop enrollment optimization system.
Three Disconnected Decisions Drive Failure:
ATOMIC
Where to place sites — reliance on tribal knowledge, not data
GEO
How to allocate across geographies — manual spreadsheet modeling
PHOTON
How to retain patients once enrolled and increase or decrease sites based upon current ongoing — reactive, not predictive
Three Products, One Integrated Pipeline — Conceived, Funded & Led by Me
ATOMIC
AI-Powered Site Selection
ML-powered site ranking using ClinicalTrials.gov enrollment data, PI publication profiles, competitive trial saturation, and internal CTMS records.
Methodology NB-GLMM, Xgboost, NLP based feature generation, SHAP explainability
Key Output Ranked site list with enrollment rate predictions per site, competitive intelligence overlay
GEO
Country & Site Optimizer
Constrained optimization engine that determines the optimal number of sites per country to minimize cost and trial duration simultaneously.
Methodology Genetic algorithm + simulated annealing, Monte Carlo uncertainty quantification
Key Output Scenario comparisons (speed vs. cost vs. balanced) with confidence intervals
AI Based explainers to help user understand all the outcomes and evaluate better solutions
PHOTON
Patient Retention Intelligence
Per-visit dropout risk scoring with explainable drivers. Identifies at-risk patients before they leave, enabling proactive site-level intervention.
Methodology Patient drop out models, bayesian methods to learn from priors to understand ongoing enrollment rate and generate triggers during risk in overall nubmer needed for enrollment
Key Output Patient risk cards with visit-level scores, protocol burden feedback loop across trials
Work Sample 01 | Scale, Adoption & Audience
Scale, Adoption & Organizational Reach
Enterprise Production Across Global Trial Operations
Why R&D Teams Operate on Incomplete Intelligence — and What It Costs
~50%
Of protocol amendments are avoidable
Each amendment costs $500K+ and delays trials 3–6 months
40hrs
To manually review one competitive landscape
Evidence buried across PDFs, registries, and literature
My Thesis
R&D researchers across Clinical, Regulatory, HEOR, and Medical Affairs all need the same intelligence — but each team builds it from scratch, every time.
I built a unified AI research platform — Design Intelligence — that serves as the shared evidence layer for all of R&D.
Four Disconnected Research Workflows:
Competitive landscape — manual registry review, no systematic tracking
Literature evidence — ad hoc searches, no reproducible synthesis
Protocol design — no institutional memory of past amendment patterns
Regulatory intelligence — label and guidance changes tracked in silos
Unique Asset Created: Built Amgen's first-of-its-kind protocol amendment database — 10,000+ protocols across all TAs. No pharma company had this before.
Work Sample 02 | Solution Architecture
My Solution & Contribution
Four Integrated Modules — One Unified Evidence Layer for R&D
TrialSight
Competitive Landscape Intelligence
Exhaustive competitive trial analysis across ClinicalTrials.gov, PubMed, and FDA registries. On-the-fly data curation with agentic AI.
Technology Agentic AI, RAG, real-time data curation, multi-source fusion
10K+ protocol amendment DB — no pharma has built this at scale
Self-Funding Model
4 business groups co-invest — revenue-generating internal product
Zero-Mandate Adoption
4x growth driven entirely by demonstrated value
R&D Leadership Visibility
Flagship example of AI creating cross-functional value
Sample Deliverables( numbers are made up )
Atomic — Site Ranking (Phase III NSCLC, illustrative)
Key Insight: MD Anderson penalized despite highest raw enrollment — 3 competing trials saturate its recruitment pool. This insight is invisible in traditional selection.
Participated in an FDA panel discussion on AI's regulatory implications and argued on Bias balance, accruyacy, and reproducviblity on some of our models
Contribvuted to Pharma Level Response to FDA on AI usage
Speakers for conference like PharmDS/ScopeX, also many times internally top tier leadership our Stance on AI and how it is going to change few workflows in Pharma and getting ready for it
Applied Research Initiatives
Led research into reproducible PDF parsing solution for accurate data extraction and developed advanced methods for biomarker extraction from complex genomic data.
Biomarker extraction from text data to understanad stratfication better in collab with Dr. Tan From Hunstan Cancer insititute ( will be published very soon, manuscript ready will be provided on request)
Peer-Reviewed Contributions
Reviewed Book on R programming by Packt - Data analysis using R Prograaming
Provided Peer to Peer reveiew to many white papers and internal papers
/Any material can be provided on request as most of the material is company related and some is still undre review, Prodcut prototypes are avaialbe on URL