STUDENT RESEARCHER  ·  AI × BIOLOGY

Decoding the undruggable.
Building the tools that find what others miss.

Hi, I'm Nimit Akhawat — a student researcher working at the intersection of AI in drug discovery, neuroscience, and biochemistry. I build computational tools and run experiments aimed at turning hard biological questions into tractable, data-driven ones.

MISSION

To make the next generation of medicines reachable — by using artificial intelligence to illuminate the dynamic, "invisible" structures of disease-driving proteins, and to connect the molecular world of chemistry with the living systems of the brain.

Explore my research → Collaborate with me
AI in Drug Discovery Intrinsically Disordered Proteins Neuroscience Optogenetics Biochemistry Machine Learning
01

Research & Projects

Computational and experimental work spanning drug discovery, neuroscience, and the chemistry that links them.

● In Progress AI in Drug Discovery

EPPS — Ensemble Persistent Pocket Scoring

A research tool for finding candidate drug-binding sites on flexible, intrinsically disordered proteins — the kind implicated in Parkinson's, Alzheimer's, and ALS that are often considered "undruggable" because they never hold a single fixed shape. Instead of one static structure, EPPS analyzes an entire ensemble of a protein's conformations and scores each residue by how persistently it lines a pocket — surfacing transient, cryptic binding sites that conventional single-structure methods miss.

The four-layer pipeline
  1. Weight each conformation by physical realism.
  2. Map per-residue pocket tendency across the weighted ensemble.
  3. Learn the structural signatures that precede pocket formation.
  4. Score — combine into a final per-residue druggability value.

Delivered as an interactive web tool: enter a protein's PDB ID and get a live, explorable analysis — 3D conformational states, a druggability landscape, and an exportable report. For proteins lacking an experimental ensemble, EPPS can generate one from a single structure via normal-mode analysis or molecular dynamics. Predictions are benchmarked against experimentally validated cryptic-pocket datasets using residue-level metrics, on the same footing as established methods in the field.

Conformational Ensembles Cryptic Pockets Normal-Mode Analysis Molecular Dynamics Druggability Scoring
● Published Neuroscience × Chemistry

Chemical Techniques in Optogenetics

A review of how synthetic and biochemical chemistry is expanding the optogenetics toolkit — from engineered photoswitchable ligands to chemically caged neurotransmitters — and how these advances give researchers finer, light-controlled command over neural circuits. The work bridges molecular design and systems neuroscience.

Optogenetics Photopharmacology Neural Circuits
Read the published article →
● Ongoing Interest Neuroscience & ML

Modeling the Brain's Dynamics

Exploring how machine-learning methods can capture brain-wave patterns and neural signaling — and how the same representational ideas that power protein modeling might illuminate neuroscience. An evolving line of inquiry rather than a single project.

Brain Waves Representation Learning Signal Modeling
02

Articles & Review Papers

Peer-reviewed publications and review articles, with summaries and links.

Review Article International Journal of Advanced Research (IJAR)

Advancing Neuroscience Through Chemical Techniques in Optogenetics

This review examines how chemistry is reshaping optogenetics — the technique of using light to control genetically targeted neurons. It surveys photoswitchable and photo-caged molecular tools, the design principles behind light-responsive ligands, and how these chemical strategies extend optogenetic control beyond classical opsins. The result is a clearer picture of how molecular-level innovation translates into more precise interrogation of neural circuits and, ultimately, new avenues for studying brain function and neurological disease.

Read article ↗ Download PDF ↓
In Preparation EPPS methods manuscript

Ensemble-Based Druggability Scoring for Intrinsically Disordered Proteins

A methods paper describing the EPPS pipeline and its benchmarking against validated cryptic-pocket datasets. Coming soon — check back for the preprint and supplementary data.

03

Main Research Goal

AI in drug discovery & development — my central focus.

Making the "undruggable" reachable.

A large share of the proteins that drive devastating diseases — many neurodegenerative disorders among them — are intrinsically disordered. They flicker between countless shapes instead of settling into one, so the standard drug-discovery playbook of "find the rigid pocket, fit a molecule into it" simply doesn't apply. These targets get labeled undruggable not because binding is impossible, but because our tools were built for a more static world.

My goal is to change which targets count as reachable. I'm developing AI-driven methods that treat a protein as the moving, breathing ensemble it actually is — learning the structural signatures that precede a transient pocket, and scoring where and when a drug might grab hold. EPPS is the first concrete step toward that vision.

🎯

Interest

The interface where machine learning meets structural biology — using data to model molecular motion, predict binding behavior, and reason about chemistry the way an experimentalist would.

🧪

Goals

Build validated, benchmarked tools that surface cryptic and transient binding sites; make them open and explorable so other researchers can act on them; and ground every prediction in experimental reality.

🔭

Vision

A future where disordered, dynamic proteins are routine drug targets — and where AI accelerates the path from a disease-linked protein to a credible therapeutic starting point.

04

Blog

Notes on neuroscience, machine learning, chemistry, and the realities of building a research career.

Neuroscience

Why "undruggable" is a statement about our tools, not the disease

A short tour of intrinsically disordered proteins and why their flexibility breaks classical drug design — and what to do about it.

Coming soon
Machine Learning

Teaching a model to see a pocket that isn't there yet

The intuition behind learning structural signatures that precede pocket formation in a conformational ensemble.

Coming soon
Chemistry

Light as a reagent: the chemistry behind optogenetics

How photoswitches and caged molecules turn a beam of light into a precise chemical command for neurons.

Coming soon
Career

Doing real research as a student

Reflections on starting independent projects early, learning in public, and finding mentors.

Coming soon
05

YouTube

Science explained — molecules, minds, and the methods in between.

▶ @nimitakhawat

Nimit Akhawat on YouTube

I make videos breaking down ideas in neuroscience, biochemistry, and AI for drug discovery — translating dense research into something you can actually follow. New explainers added regularly.

▶ Watch on YouTube Subscribe
07

Resume

Academic achievements and a downloadable CV.

At a glance

  • Focus: AI in drug discovery, neuroscience, and biochemistry.
  • Flagship project: EPPS — ensemble-based druggability scoring for intrinsically disordered proteins.
  • Publication: Review article on chemical techniques in optogenetics (IJAR).
  • Science communication: YouTube channel translating research for a wider audience.
  • Skills: Computational structural biology, machine learning, molecular dynamics, scientific writing.
📄

Get the full CV with detailed coursework, projects, and achievements.

Download Resume ↓ Place your PDF at assets/resume/Nimit-Akhawat-Resume.pdf
08

Contact & Collaborate

Open to collaborations, mentorship, and conversations with fellow researchers. Let's build something.

Your message is saved securely to my inbox. I usually reply within a few days.

Reach me directly

✉ hello@nimitakhawat.com

For research collaborations, speaking, or mentorship inquiries.

Find me online

Update these links in index.html with your profiles.