Medicinal chemistry | chemical biology | fluorescent probes | transporter biology | spectroscopy | QSAR | molecular AI | open science
Recognition: UCSF PBBR Independent Research Award · Harvard Business School Foundry Bootcamp · ACS Postdoc to Faculty · Sigma Xi Full Member · Excellence in Doctoral Research Award · NCCR Bioinspired / EYCN & Swiss Academy Sciences Special Sponsor Award
I am building JAI Lab, an independent open research initiative for molecular discovery.
The thesis is simple:
The next leap in molecular AI will come not only from larger models, but from better scientific infrastructure: cleaner datasets, stronger benchmarks, reproducible software, and experiments that machines can actually read.
I am connecting synthetic chemistry, fluorescent probes, transporter biology, analytical chemistry, molecular modeling, machine learning, and open-source scientific software.
| Signal | Evidence |
|---|---|
| Scientific base | Medicinal chemistry, synthetic organic chemistry, fluorescent probes, chemical biology, ion transporter pharmacology |
| Research output | 8 peer-reviewed publications; h-index 6; 75 citations as listed in my June 2026 CV |
| Independent funding | UCSF PBBR Postdoctoral Independent Research Award for a xylazine fluorescent sensor project |
| Recognition portfolio | Harvard Business School Foundry Bootcamp; ACS Postdoc to Faculty Workshop; Sigma Xi Full Member; Excellence in Doctoral Research Award; NCCR Bioinspired / Swiss Academy Special Sponsor Award |
| Open-source direction | Public GitHub projects spanning sensing, spectroscopy, QSAR, docking, paper organization, and molecular datasets |
The recognition portfolio spans independent biomedical research funding, faculty-track preparation, scientific entrepreneurship, doctoral excellence, and international chemistry-program recognition.
| Recognition | Year | Signal |
|---|---|---|
| UCSF PBBR Postdoctoral Independent Research Award | 2024 | Independent high-risk, high-reward biomedical research funding for a xylazine fluorescent sensor project |
| Harvard Business School Foundry Bootcamp | 2026 | Entrepreneurship, venture development, market validation, and investor-pitch training |
| ACS Postdoc to Faculty Workshop | 2025 | Competitive faculty-transition mentoring and career-development program |
| Sigma Xi Full Member | 2026 | Induction into The Scientific Research Honor Society |
| Excellence in Doctoral Research Award | 2022 | Doctoral research recognition from the School of Graduate Studies, Ari'el University |
| Special Sponsor Award, NCCR Bioinspired / Swiss Academy of Sciences | 2022 | International recognition at the European Young Chemists Meeting, Fribourg, Switzerland |
| Area | Repository | What it does |
|---|---|---|
| Molecular sensing | SensorGenome | AI platform for molecular sensing datasets, benchmark tasks, active learning, and autonomous sensor discovery |
| Spectroscopy and analytical chemistry | NarcoticSense-AI | Open-source AI platform for spectroscopy, chemometrics, and narcotic sensing research |
| Xylazine analytics | AZAI | AI-driven xylazine analytics and innovation for emerging adulterant detection |
| Fluorophore data | BimaneDB | Open bimane fluorescent dye database and preliminary QSAR modeling |
| Transporter drug discovery | pendrin-qsar / pendrin-hybrid | Ligand-based and structure-based modeling of pendrin inhibitors |
| CFTR molecular modeling | cftr-qsar | QSAR modeling of CFTR potentiators using RDKit and ML |
| Research tooling | Paper-Organizer | AI-powered research paper organizer for scientific literature workflows |
| Personal site | drjoykarmakar.github.io | Personal scientific website and research portfolio |
| Program | Core question | Public artifact |
|---|---|---|
| SensorGenome | How do we make molecular sensing experiments machine-readable and benchmarkable? | Dataset schemas, benchmark cards, active-learning workflows |
| DyeMind | Can AI accelerate fluorophore discovery? | Predictive models, generative design ideas, fluorophore datasets |
| NarcoticSense AI | Can spectroscopy and chemometrics provide useful chemical intelligence? | Spectral preprocessing, classification, analysis pipelines |
| AZAI | How can molecular sensing support analytics for xylazine and emerging adulterants? | Xylazine-centered computational and sensing workflows |
| Molecular Discovery Suite | How can QSAR, docking, SAR, and ADMET support transporter-focused drug discovery? | Pendrin, PAT1, CFTR, and related molecular modeling workflows |
| Open Science Tools | How do we make research reusable by default? | Templates, dataset cards, benchmark cards, documentation systems |
My research training is experimental first, computationally expanding, and infrastructure-focused.
|
|
At UCSF, I worked on small-molecule modulators of ion transporters including Pendrin, PAT1, and CFTR-relevant chemistry. My postdoctoral research included SAR-driven optimization of Pendrin inhibitors and PAT1 inhibitors, with sub-micromolar to low-micromolar leads and in vivo validation in collaborative biological models.
My doctoral research focused on novel bimane derivatives, luminescent chemical tools, and sensing applications. This foundation now informs DyeMind, BimaneDB, and molecular sensing projects aimed at turning fluorescent probe discovery into a more data-rich and AI-ready field.
Through the UCSF PBBR award project and the AZAI/NarcoticSense AI direction, I am exploring how fluorescent probes, spectroscopy, chemometrics, and molecular AI can support chemical intelligence for xylazine and related emerging adulterants.
I am building public tools that make chemistry easier to reproduce: curated molecular datasets, benchmark templates, QSAR notebooks, spectroscopy pipelines, and repository standards that make scientific software more useful to other researchers.
Peer-reviewed publications: 8
Google Scholar metrics listed in CV: h-index 6, 75 citations
Selected journals: European Journal of Medicinal Chemistry, RSC Medicinal Chemistry, Chemical Communications, Frontiers in Chemistry, Synlett, Israel Journal of Chemistry, Inorganica Chimica Acta
Selected publications:
- High potency 3-carboxy-2-methylbenzofuran pendrin inhibitors as novel diuretics. European Journal of Medicinal Chemistry (2024).
- Selective isoxazolopyrimidine PAT1 (SLC26A6) inhibitors for therapy of intestinal disorders. RSC Medicinal Chemistry (2023).
- A dipodal bimane-ditriazole-diCu(II) complex serves as ultrasensitive water sensor. Chemical Communications (2022).
- Highly sensitive water detection through reversible fluorescence changes in a syn-bimane based boronic acid derivative. Frontiers in Chemistry (2022).
For the full list, see Google Scholar.
| Principle | Meaning |
|---|---|
| Better data beats bigger claims | Molecular AI is only as useful as the experimental data behind it. |
| Benchmarks create accountability | Progress should be measured through transparent tasks, not vague demos. |
| Experiments should be machine-readable | Protocols, conditions, controls, uncertainty, and outcomes belong in structured datasets. |
| Models must return to chemistry | Predictions need experimental context, synthetic feasibility, and validation. |
| Open infrastructure compounds | Reusable datasets, templates, and software help the entire field move faster. |
I organize repositories as scientific products, not storage folders. A strong JAI Lab repository should include:
- a clear scientific question
- a reproducible environment
- examples that run quickly
- dataset or benchmark cards when relevant
- citation metadata
- roadmap and known limitations
- contribution instructions
- Convert SensorGenome into the flagship public standard for molecular sensing datasets and benchmark cards
- Release polished dataset-card and benchmark-card templates for molecular AI projects
- Standardize repository READMEs across SensorGenome, AZAI, NarcoticSense AI, BimaneDB, pendrin, and CFTR projects
- Expand DyeMind into a dedicated fluorescent-probe AI workspace
- Add reproducible notebooks for QSAR, docking, spectroscopy, and active-learning workflows
- Create a documentation site for JAI Lab projects, tutorials, and research notes
- Link publications to code, datasets, and reproducibility artifacts whenever possible
Chemistry: ChemDraw, MestreNova, Mercury, ACD/NMR, HPLC, LC-MS, HRMS, NMR, UV-Vis, fluorescence spectroscopy
Cheminformatics and modeling: RDKit, AutoDock, AutoDock Vina, Maestro, SwissADME, ADMETlab, ORCA, Avogadro
AI and data: Python, PyTorch, TensorFlow, scikit-learn, pandas, NumPy, QSAR, molecular descriptors, active learning, uncertainty-aware modeling
Open science: GitHub Actions, Docker, reproducible notebooks, dataset cards, benchmark cards, documentation systems
I am interested in collaborations across molecular sensing, fluorescent probes, transporter chemical biology, medicinal chemistry, spectroscopy, analytical chemistry, AI/ML for chemistry, drug discovery, open datasets, benchmark development, and reproducible scientific software.
Connect through joykarmakar.com, LinkedIn, Bluesky, or GitHub.
