AI/ML for Drug Discovery
Predictive toxicity models (TLR8 and beyond) that discover novel sequences in silico, validated by wet lab data, progressing toward clinical trials.
AI Champion at GSK, representing 3,000+ scientists. I bridge deep biological understanding with machine learning, automation, and data infrastructure to accelerate oligonucleotide drug discovery -- from in silico prediction to clinical progression.
I operate at the intersection where experimental biology generates data, data becomes predictive models, and models drive real drug discovery decisions. I speak fluently with biologists about assay design, with ML engineers about feature engineering, and with leadership about what the science means for the pipeline.
Click any card to see the technical detail underneath.
Predictive toxicity models (TLR8 and beyond) that discover novel sequences in silico, validated by wet lab data, progressing toward clinical trials.
Automated hit-picking across multiple toxicity liabilities and AI agents for bench scientists -- combining lab automation (Tecan, Hamilton, Integra) with intelligent workflow tools.
Reproducible Python pipelines for plate reader outputs, NGS data, QC, normalization, and reporting that turn messy experimental data into model-ready datasets.
AI Champion at GSK for 3,000+ scientists, building AI agents for bench-level researchers, driving enterprise-wide responsible innovation and standardization.
Research, AI/ML, automation, scientific communication, and leadership -- grouped into deeper project pages.
Appointed AI Champion at GSK, driving responsible AI adoption, building AI agents for bench scientists, and leading an AI Task Force that surfaces, packages, and scales AI use cases across a 3,000+ person R&D organization.
Building predictive models for ASO and siRNA safety endpoints -- TLR8, complement activation, GPVI, AGO2 binding, and RNA accessibility. The TLR8 model enabled in silico discovery of novel sequences in a portfolio project with clinical prospects.
Combining automated multi-liability hit-picking with AI agents built for bench scientists. From days of manual data processing to minutes of automated execution, with Claude-powered tools for standardization and optimization across workflows.
Independent assay development across complement activation, caspase, RNA accessibility, and free uptake. Running 25+ NGS screening campaigns, deep cell biology fluency, and novel high-throughput assay design for oligonucleotide safety and efficacy.
A 12-week self-directed curriculum translating machine learning into language bench scientists already understand.
Neurotoxicity and cardiotoxicity assays, imaging workflows, and Python-driven analysis automation.
Experimental design, immunology, and translational interpretation at UC San Diego.
Using AI as a structured scientific thinking partner for experiment planning and data interpretation.
Clinical documentation and science communication for high-stakes regulated contexts.
Science storytelling, editorial leadership, and community-building at scale.
Making experimental biotechnology understandable and teachable.
Hands-on liquid handling across Tecan Fluent, Hamilton STAR, and Integra platforms. Tecan FluentControl Intermediate and Advanced certified. Translating manual assays into scalable automated workflows.
Connecting students with biotechnology and engineering industry leaders.
Building platforms for biotechnology ideas, teams, and execution.
Clinical lab exposure and genetics workflows.
1000+ manuscripts prepared across biology and life sciences.
A career built on the principle that the best AI in drug discovery comes from people who deeply understand the science.
Expanding into business strategy and leadership to complement scientific and technical depth.
Appointed AI Champion at GSK for 3,000+ scientists. Building predictive toxicity models, automated hit-picking platforms, and AI agents. Leading AI Task Force for enterprise-wide adoption.
RNA therapeutics safety screening, high-content imaging, Python analysis pipelines. Rapidly learned new assay systems and built automation scripts that streamlined team workflows.
Thesis on Type I interferon signaling in tumor progression and immune cell function. Teaching assistant for CRISPR-Cas9 and scientific writing. Tata Scholar.
Clinical study report drafting and lean writing initiatives in regulated clinical development.
1000+ manuscripts prepared. Built fluency in publication standards and scientific communication at scale.
Ranked 2nd in department. Founded The BioTalk Magazine and Bio-Entrepreneurship Competition. Branch Ambassador Award.
My writing spans mechanistic biology, oligonucleotide safety, AI strategy, and public-facing science communication.
A public argument for why scientists should lead AI adoption, not resist it.
First-author peer-reviewedCytokine and Growth Factor Reviews, 2025.
Full ProfileComplete publication record and citation profile.
Science Commentary -- The Hindu, 2025Analysis of India's biotech sector trajectory and structural challenges.
AI & Biomanufacturing -- The Hindu, 2025Policy analysis on AI integration in India's biomanufacturing sector.
Presented WorkPresented across oligonucleotide therapeutics and gene therapy forums.
PublicationInternational Journal of Research and Analytical Reviews.
Review PaperInternational Journal of Advance Research, Ideas and Innovations in Technology.
Learning SystemSelf-built AI/ML curriculum translating machine learning into molecular biology language.
Deepakshi was a pleasure to work with. She works independently and can search the literature, design experiments, and follow protocols. She has the potential to become an outstanding scientist.
Dr. Jack Bui, Director, UC San Diego
Despite initially having no prior experience in these specific assays, Deepakshi quickly acquired the necessary skills to conduct experiments independently. Her efforts were instrumental in developing automation scripts that streamlined our data analysis pipeline.
Rafael Renteria, Principal Neuroscientist, Creyon Bio
After being given a question regarding assay optimization, she independently conceptualized, planned, executed, analyzed, and interpreted experiments. The experiments she planned were robust and well controlled.
Tim Nicholson-Shaw, Sr. Scientist and RNA Biologist
Deepakshi is diligent, creative and an excellent communicator. She learned Python skills to analyze experimental data and is passionate as a scientist.
Swagatam Mukhopadhyay, Chief Scientific Officer, Creyon Bio
She is an excellent communicator with great attention to detail. Her leadership skills and can-do attitude make her a valuable addition to any team.
Chinmayi Kashyap, UC San Diego collaborator
Deepakshi channelized her passion, leadership qualities, and management skills toward building a strong scientific community of biotechnology enthusiasts.
Nikhil Patil, Scientist and BioTalk Magazine advisor
The same energy that drives my scientific work shapes how I lead, learn, and collaborate outside of it.
I am always interested in conversations about AI-driven drug discovery, scientific automation, and the future of computational biology.