A prediction is not an experiment.
AI can prioritize hypotheses, molecules and experiments. It does not turn a computed score into measured evidence. BayesPharma Labs is designed to keep those evidence classes visibly separate.
About / BayesPharma Labs
Build the lab
that learns.
BayesPharma Labs is an India-first therapeutics R&D company being built around a simple scientific idea: predictions become useful when they meet experiments, experiments become powerful when they return structured evidence, and evidence should shape the next decision.
INTERNAL THERAPEUTICS R&D DIGITAL TWIN · LIVE PHYSICAL LAB · STAGED DEVELOPMENT
BAUTONOMY CORE
DESIGNAI + models
MAKEsynthesis
TESTexperiments
LEARNevidence
DECIDEhuman gate
01 / WHY WE EXIST
Drug discovery is not short of predictions.
It is short of fast, trustworthy loops between a scientific question and experimental truth. We are designing BayesPharma Labs around that gap.
Every machine action should leave provenance.
Robotics only matters scientifically when the system records what happened, under which protocol, with which material, instrument, model version and uncertainty.
Autonomy should be bounded, not mystical.
Software can plan and robots can execute repeatable work. Consequential programme gates remain reviewable and accountable to qualified scientific judgment.
02 / HOW WE THINK
A discovery system should close the loop.
The goal is not to automate isolated tasks. It is to connect intent, execution, evidence and the next experiment without losing scientific context at each handoff.
01 / FRAMEQuestionDefine target, therapeutic objective, constraints and success criteria.
02 / DESIGNHypothesisGenerate and rank molecules or experiments using models and evidence.
03 / MAKEExecutionTranslate selected work into synthesis and machine-addressable protocols.
04 / TESTMeasurementCapture assay, analytical and quality-control outputs with provenance.
05 / LEARNEvidenceUpdate models and uncertainty without erasing negative or conflicting results.
06 / GATEDecisionAdvance, redesign, stop or run the next experiment through explicit review.
COMPUTEDPREDICTEDSIMULATEDMEASUREDREVIEWED
03 / WHY INDIA
From pharmacy of the world to a birthplace of original medicines.
India already has chemistry, manufacturing depth, clinical talent and engineering scale. The opportunity is to connect them with original asset creation and faster experimental learning.
BayesPharma Labs is being built with an India-first lens: choose important therapeutic problems, own the scientific hypothesis, build the evidence chain, and progressively internalize the experimental capabilities that create differentiated learning.
That means starting software-first where appropriate, qualifying each new physical capability instead of pretending it already exists, and using partnerships and CRO execution where they are scientifically and economically sensible.
Original assetsOwn the question, candidate and rationale.
Evidence disciplineKeep provenance and uncertainty visible.
Physical AIAdd robotics where repeatability earns it.
Global ambitionBuild for medicines, not demos.
FOUNDER / 001HYDERABAD · INDIAPHARMA × AI × AUTONOMOUS SCIENCE
GU
04 / FOUNDER
Gunda
Upendar Rao.
Founder · BayesPharma Labs
Gunda Upendar Rao is a Hyderabad-based pharmaceutical-science-trained founder and sci-tech builder. He pursued his pharmacy graduation at University College of Pharmaceutical Sciences, Kakatiya University, and later advanced pharmaceutical-science training at the National Institute of Pharmaceutical Education and Research (NIPER), Mohali, from 2010–2012.
His founder journey spans pharmaceutical intelligence, healthcare, education, games, scientific software and frontier-science ventures. That path includes NextDNA Edutech, CuriousAtoms, LearnPlay, KalpanaSpace and other experimental products built around a recurring question: how can difficult knowledge become more useful, interactive and actionable?
His entrepreneurial work has also crossed education, learning games and health technology: NextDNA Edutech, CuriousAtoms, LearnPlay, FutureMinds Lab and SEMwell / DawaiSafe. Across these projects, the common thread is building practical technology around science, learning, experimentation and access.
At BayesPharma Labs, he is pursuing a practical version of autonomous science: use AI to frame and prioritize better experiments, use automation to execute repeatable work, keep experimental evidence inspectable, and preserve human accountability at consequential scientific gates.
UCPSC · KAKATIYA UNIVERSITYNIPER MOHALI · 2010–2012SCI-TECH FOUNDERPHARMA × TECHNOLOGYHYDERABAD, INDIA
“Fallibilistic and a Fun-driven Sci-Tech Adventurer.”
05 / MATURITY, NOT MYTH
Say what exists.
Build what comes next.
What exists today
01Digital Twin and software-defined discovery workflows.
02India-first therapeutic programme prioritization and candidate evidence architecture.
03Computational design, docking, developability, synthesis-planning and evidence workflows.
04Human scientific review and explicit evidence-state boundaries.
What we are building toward
01Qualified robotic workcells for repeatable experimental execution.
02Machine-readable synthesis, assay and analytical handoffs.
03Closed-loop Design → Make → Test → Learn cycles with bounded autonomy.
04Experimental data that improves the next scientific decision—not merely the next dashboard.
Operating boundary: BayesPharma Labs does not present concept visualizations or Digital Twin states as measured wet-lab evidence. Physical laboratory capability is being developed in stages and should be judged by qualified, reproducible execution as it comes online.
06 / MISSION
The goal isn't more experiments.It's better medicines.
Build faster learning loops. Keep evidence honest. Automate what can be made repeatable. Keep scientific accountability visible.
Explore BayesPharma Labs →