Profile banner
Sarvesh Mehta

Sarvesh Mehta

ML researcher and engineer — chemical ML and trading ops

India
linkedin
Joined October 2025
Favicon imagein.linkedin.com
Favicon imagescholar.google.com
Favicon imageresearchgate.net

Network

1.1K connections
🎯
AI Product Leaders
🔧
Software Engineers
📈
High-Frequency Trading Professionals
💻
IIIT Hyderabad Big Tech SWEs
🎓
IIT/IIIT Network
🤖
AI/ML Professionals
☁️
...

Summary

Applied researcher in chemical machine learning: while at IIIT Hyderabad Sarvesh co‑authored and led multiple projects (MEMES, MO‑MEMES, and a JPCL paper) that apply Bayesian optimization, graph neural networks and reinforcement learning to problems in virtual screening and spectral inversion, and helped produce public code for MO‑MEMES. nih+3
Open‑source contributor and community participant: Sarvesh has contributed to open projects (Google Summer of Code with CiviCRM in 2019 and 2020) and produced code connected to his research, indicating experience both in academic research and shipping usable code. withgoogle+2
Transitioned into trading operations and startups: professional profile listings indicate a move from IIIT Hyderabad research into trading‑operations engineering at Flow Traders (graduate program then Trading Operations Engineer) and subsequently co‑founding a stealth startup in 2025, showing a shift from academic research to production engineering and entrepreneurship. linkedin

Work

Education

Projects

Writing

MO-MEMES: A method for accelerating virtual screening using multi‑objective Bayesian optimization

January 1, 2022

Extends MEMES to a multi‑objective setting to identify molecules on the Pareto front across multiple properties; includes code (MO‑MEMES repository).

Favicon imagepmc.ncbi.nlm.nih.gov

Deep Reinforcement Learning for Molecular Inverse Problem of NMR Spectra to Molecular Structure

January 1, 2022

Describes an RL + graph convolution approach to infer molecular structure from NMR spectra; demonstrates strong top‑k recovery for small molecules.

Favicon imageweb2py.iiit.ac.in

MEMES: Machine learning framework for Enhanced MolEcular Screening

January 1, 2021

A Bayesian‑optimization‑based machine‑learning framework designed to efficiently sample chemical space for high‑throughput virtual screening; demonstrates identifying a large fraction of top hits while evaluating only a small portion of libraries.

Favicon imagepmc.ncbi.nlm.nih.gov

Hobbies

Open‑source contributions (Google Summer of Code projects and public code repositories). withgoogle+2

Research in computational chemistry / machine learning (publications and academic projects). nih+1