Sanatan Upmanyu

• — Co-founded the company and own the AI/data-science function end to end — technical strategy, architecture, hiring, and delivery across the product portfolio.
• — Direct build and deployment of SequenceFlow ResearchNexus and Clinical Navigator: an integrated stack spanning knowledge graph, clinical-trial prediction, literature analysis, patient digital twins, TPP and TOP, with 12 ML models in production on PostgreSQL + Neo4j and self-hosted LLM APIs.
• — Set the AI/ML agenda for early innovation-signal detection across emerging targets, novel therapeutics, and healthcare solutions — and translate it into the commercial product roadmap.
• — Steer the Asset Lifecycle Manager and Portfolio Management platforms, extending the company's reach from discovery signals into portfolio decision-making.
• — Lead delivery of AI engagements for top-tier global pharmaceutical companies (client names confidential).

• — Site head for the Basel operation and owner of the data-sciences function — team of 5 data scientists, hiring, technical direction, and site-level delivery accountability.
• — Own the AlphaMeld Operating System end to end — a custom UI layer (Knowledge Graph, ChatAlphaMeld, Bioinformatics, Literature Analysis, AlphaClinMeld) over 12 ML models — compressing hypothesis-to-insight cycles from months to weeks.
• — Delivered AlphaCompass portfolio intelligence: adopted by two business-development teams within six months, standardizing evidence-based indication prioritization across 200 active programs.
• — Directed the ChatAlphaMeld benchmark programme — 97.34% accuracy on a 188-question biomedical benchmark, outperforming leading general-purpose frontier models and delivering the step-by-step rationales complex clinical questions demand.
• — Anchor platform delivery and support for the publicly announced collaborations that validate AlphaMeld commercially: Kyowa Kirin (multi-year strategic partnership; subscription-and-support relationship spanning five years, with an AI-identified program validated in preclinical proof-of-concept), Shionogi (multi-target drug discovery, worth up to $200M per program in upfronts, milestones, and royalties), GSK Consumer Healthcare (three-year AI innovation collaboration), and Ono Pharmaceutical (2024 target discovery).
• — Act as senior data-science advisor to pharma and biotech clients, applying molecular-biology depth to MOA analysis and target-viability calls that shape client program decisions.

• — Own technology strategy, architecture, and delivery for the company's entire digital product portfolio — concept to production — and the engineering function behind it.
• — Conceptualized and built QuriousRI, a secure role-based regulatory-intelligence platform covering pharmaceuticals, biologics, biosimilars, and generics across 12+ agencies (FDA, EMA, PMDA, NMPA, Health Canada, MHRA, Swissmedic, and more).
• — Extended QuriousRI into CTD dossier preparation, driving AI-assisted authoring across Modules 2 through 5 — and scoping Module 1 out deliberately, as country- and company-specific content that does not transfer reliably. Reproducibility tested and repeated to date on Module 2, the Quality Overall Summary, 3.2.P.1, and 3.2.P.3.3, with section coverage expanding.
• — Built per-module, per-section regulatory gap assessment for market expansion: for a drug already approved in one market, the platform resolves what a target market requires for that drug type and where those requirements diverge from the original filing — converting a manual, consultant-heavy exercise into a repeatable product capability.
• — Architected an AI assistant that mines global regulatory updates and auto-drafts submission dossiers for first-in-class device filings, opening a new non-pharma revenue channel.
• — Direct enterprise data-science and AI product development on Agile, holding delivery to FDA, EMA, and ICH standards across the drug-development lifecycle.

• — Board-level advisor on AI-driven strategy for early-stage drug discovery and translational research.
• — Guide regulatory-aware innovation across the company's digital and data-science agenda.

• — Co-founded the firm and built the data-sciences practice from zero — team, methodology, delivery model, and client base.
• — Advised pharma and biotech leadership on de novo, reformulation, repurposing, and drug-rescue strategy grounded in real-world data.
• — Built statistical and machine-learning predictive models over RWD using NLP; led therapeutic target assessment and MOA analysis for client programs.

• — Set short- and long-term strategy for the data-science and AI organization.
• — Owned enterprise-scale data-science products end to end across the drug-development lifecycle, delivered on Agile.
• — Conceptualized and shipped novel solutions spanning discovery through clinical development.

• — Directed drug-repurposing and novel-target hypothesis generation using big-data and statistical methods.
• — Built and mentored the data-science team, owning delivery for US-based pharma and biotech clients.
• — Advised clients on de novo, reformulation, repurposing, and drug-rescue strategy, plus surveys and future-opportunity mapping using real-world data.
• — Led strategy consulting and project management, providing the scientific backing for business development.

• — Ran siRNA targeted gene silencing for target-based drug discovery.
• — Cloning, expression, purification, and activity analysis of human kinases, cytochrome P450 proteins, GPCRs, and bacterial enzymes — supporting enzyme- and cell-based NCE screening assays across mammalian, baculovirus, and Dictyostelium expression systems.

End-to-end AI stack — Knowledge Graph, ChatAlphaMeld, Bioinformatics, Literature Analysis, and AlphaClinMeld over 12 ML models (NER, relation extraction, evidence scoring, digital-twin trial predictors, landscape assessment) on PostgreSQL + Neo4j with self-hosted LLM APIs.

Secure role-based regulatory platform for pharma and biotech affairs teams, built from concept as CTO. Beyond search and comparison of global approvals, it drives CTD dossier authoring and per-section gap assessment across Modules 2 through 5 — Module 1 deliberately excluded as non-transferable country- and company-specific content — telling a sponsor exactly how a target market's requirements diverge from an existing filing. Reproducibility tested and repeated to date on Module 2, the Quality Overall Summary, 3.2.P.1, and 3.2.P.3.3.

Clinical-trial–centric SaaS module on AlphaMeld® graph analytics; benchmarks pipeline assets and quantifies comparative risk. Adopted by two BD teams within six months, standardizing evidence-based indication-prioritization across 200 active programs.

End-to-end AI stack combining Knowledge Graph, clinical-trial prediction, literature analysis, patient digital twins, TPP and TOP, with 12 ML models on PostgreSQL + Neo4j and self-hosted LLM APIs.

Guided hypothesis-generation platform on the AlphaMeld Knowledge Graph, supporting disease-first, target-first, indication-expansion, and combination-therapy entry points through a project → exploration → agent scoring → ranking workflow. Multi-tenant SaaS with SSO, RBAC, and audit logging.

Digital-twin framework over a 23,000+ patient, 200+ center systemic sclerosis registry. A PyTorch deep/generative stack produces prognostic scores and synthetic placebo trajectories — an ML-generated digital control arm that lifts trial power and cuts sample size, aligned with FDA digital-twin and EMA/ICH MIDD guidance.

Two agentic workflow engines fine-tuned on biomedical tasks; outperformed leading general-purpose frontier LLMs; delivered richer step-by-step rationales critical for complex clinical questions.