About Work

~/hello — welcome to my corner of the internet

Deep
Kotadia

M.Tech Chemical Engineering · IIT Madras
Scientific ML | Local LLM Systems | Cheminformatics | Process Engineering

Hi, I'm Deep — a postgrad chemical engineering student at IIT Madras, hooked on building at the intersection of science and intelligent systems.

My work focuses on Scientific Machine Learning — from predicting adsorption energies of catalysts using ML models to building LLM-powered agentic pipelines that mine literature at scale for data curation. I'm particularly excited about how local LLM systems and automated scientific workflows can accelerate research that would otherwise take years.

Outside of research, I have a background in process engineering (Zentiva, TITLIS), graphic design leadership, and a published patent. I'm driven by curiosity, collaboration, and the belief that sustainable energy solutions live at the intersection of data and domain expertise.

Deep Kotadia
Chennai, India

Current Endeavors

What I'm Working On

Actively Exploring

  • Building an agentic LLM pipeline for automated curation of catalytic depolymerisation datasets from scientific literature
  • Designing local LLM inference systems optimized for scientific Q&A and data extraction tasks

Learning

  • Studying advanced transformer architectures and their application to molecular data
  • Deep-diving into cheminformatics workflows using RDKit and molecular fingerprinting for property prediction

Beyond the Lab

A Glimpse Outside Work

Interests

When I'm not busy with code or running experiments, you'll find me at the piano, behind a camera, or out on a long cycle ride. This year I picked up meditation — small daily rituals that quietly anchor everything else. And I love exploring places — hill stations, coastal towns, temple cities.

Piano Photography Cycling Meditation Travel
Places I've Wandered
Shimla Manali Ooty Coorg Kodaikanal Tirumala Rameshwaram Pondicherry Vaishno Devi Igatpuri

Currently

Building Designing local LLM inference systems optimized for scientific Q&A and data extraction tasks.
Listening Gujarati folk music — loud, raw, and turned all the way up

Gujarati at heart · always up for a chai over a long chat — half my best ideas land somewhere between sips.

Through My Lens
At the piano
Photograph
Photograph
Cycling
Photograph
Photograph
Photograph
Photograph
Photograph
Photograph
Photograph
Photograph
Photograph
Photograph
Photograph

Tech Arsenal

Skills & Tools

Scientific ML & Data Science
ML ModelsXGBoost, SVR, Random Forest, Neural Networks, PINNs
Deep LearningPyTorch, TensorFlow, Keras, Hyperparameter Tuning
LibrariesScikit-learn, NumPy, Pandas, Matplotlib, Seaborn
MLOpsFastAPI, Docker, Git, GitHub
LLM & Agentic Systems
FrameworksLangChain, n8n Workflows
Local LLMOllama, Unsloth, vLLM, LM Studio
CapabilitiesLocal LLM Design, Prompt Engineering, Agentic Pipelines
ToolsHermes Agent, VS Code, Ollama-style inference
Cheminformatics & Chemistry
LibrariesRDKit, SMILES, Morgan Fingerprints
MethodsTanimoto Similarity, Molecular Descriptors
TasksLiterature Mining, Dataset Curation, Property Prediction
Process Engineering
SimulationAspen Plus, Diagrams.net, DWSIM
SafetyHAZOP, What-If Analysis, Process Safety
LanguagesPython, MATLAB, SQL

Education

Academic Foundations

M.Tech Jul 2024 — Jun 2026

Chemical Engineering

Indian Institute of Technology, Madras
8.71
CGPA

M.Tech Thesis · Prof. Jithin John Varghese & Prof. Tarak Patra

Curated Dataset on Catalytic Depolymerization of Polyethylene Using Large Language Models.

  • Developed an end-to-end AI-driven literature mining pipeline using Gemma4:e4b (Local LLM) to automate extraction of catalytic pyrolysis data from 736+ scientific publications
  • Curated a structured dataset of 559 experimental datapoints from 73 research papers — covering 120+ catalysts, 6 reactor configurations, and operating conditions for catalytic polyethylene pyrolysis
  • Built and optimized ML models (XGBoost, Random Forest, SVR) using Optuna-TPE for hydrocarbon product yield prediction — R² up to 0.78 for gas yield estimation
  • Engineered a fully local GPU-accelerated scientific workflow integrating OCR, PDF parsing, prompt engineering, and schema validation — significantly reducing manual data curation effort

Teaching Assistantship

CH5650
Molecular Data Science & Informatics
Jan 2026 — May 2026 · Prof. Tarak Kumar Patra · Delivered ANN tutorials & cheminformatics workflows (RDKit, SMILES) to 70+ students
CH3010
Chemical Reaction Engineering
Jul — Dec 2025 · Prof. Jithin John Varghese · Designed Python/Jupyter tutorials on reactor design and ODE-based kinetics for 100+ students

Key Coursework

Molecular Data Science Informatics Application of Machine Learning in Reaction Engineering Computational Methods in Catalysis Artificial Intelligence in Chemistry Process Simulation Laboratory Chemical Reactor Theory Advanced Chemical Engineering Thermodynamics

Activities

Data Science Intern at Rapidrug Pvt. Ltd. (Nov 2024 — Jan 2025)

Graphic & Web Design Intern at MeDow (Jan 2025 — March 2025)

Website Design Coordinator at CARES - Catalysis Reaction Engineering Symposium, IIT Madras (Sep 2025)

B.Tech Aug 2019 — May 2023

Chemical Engineering

Pandit Deendayal Energy University, Gandhinagar
8.57
CGPA

Major Project

Process Simulation of Ethyl Lactate Production in DWSIM — A Green Solvent.

  • Simulated and optimized ethyl lactate production via reactive distillation in DWSIM — 96.13% lactic acid conversion and 95.7% yield
  • Designed a flowsheet integrating fermentation, reactive distillation, and product purification — 234 g/h production rate, 111.41 kJ/g energy consumption
  • Applied NRTL property model for non-ideal mixtures to optimize operating parameters for high-purity output

Relevant Coursework

Material Science & Engineering Nanotechnology & Energy Storage Pharmaceutical Technology Polymer Science Technology Process Intensification Computer Aided Process Design & Laboratory Process Modelling & Optimization Laboratory Process Equipment Design & Laboratory Transport Phenomena IoT for Industries Industry 4.0

Leadership & Activities

  • Head of Graphic Design · Indian Institute of Chemical Engineers (IIChE) Student Chapter
  • Head of Graphic Design · Cretus — The Robotics and Automation Club
  • Vehicle Dynamics (Brakes) · Team Ragnar — Formula Student Team, PDEU

Experience

Career Timeline

Nov 2024 - Jan 2025
Data Science Intern
Rapidrug Pvt. Ltd.  ·  Chennai, India
  • Developed an AI-driven veterinary drug safety forecasting system integrating time-series analytics and molecular intelligence.
  • Engineered molecular descriptors using RDKit and combined cheminformatics workflows with predictive modeling.
  • Designed agentic AI workflows for automated drug trend analysis and veterinary pharmaceutical insights.
  • Applied forecasting models and exploratory ML techniques for companion animal therapeutic analytics.
Jan 2023 — Jun 2023
Process Technology Engineer
Zentiva Pvt. Ltd.  ·  Ankleshwar, India
  • Achieved 88% yield by scaling up Grignard reagent formation and optimizing multi-step target API synthesis.
  • Ensured 93% yield in intermediate synthesis through robust phase-transfer catalysis using NaH and TBAB.
  • Developed process flowsheets for Grignard reagent, intermediate, and API coupling to enable efficient production.
  • Performed process safety assessments using What-If and HAZOP approaches, addressing runaway reaction, cooling loss, overcharging, and moisture risks.
May 2022 — Jul 2022
Process Design Engineer (Intern)
TITLIS Projects & Engineering Pvt. Ltd.  ·  Vadodara, India
  • Modeled a 3-stage silicon extraction process in Aspen Plus, running sensitivity analyses and "what-if" optimization studies on reactor conditions (T, P, composition)
  • Evaluated conversion efficiencies across process units, achieving 85% SiO₂ to metallurgical-grade silicon conversion and 90% trichlorosilane selectivity while quantifying residual impurity profiles including Fe, Al, Ti, P, and B.

Contact

Let's Connect

The best way to reach me is through email. I'm always open to research collaborations, project discussions, or just a conversation about ML and chemical engineering.