Experience

Digital Twin Engineer (optimisation and AI engineering) · Gradyent · 12/2025 – present

Gradyent’s digital-twin platform recommends the optimal operation of industrial and district-heating networks. I work on its optimisation software and build AI tooling.

• Optimisation: I helped deliver over 3x faster compute and ~97% lower peak memory for a live optimisation service, and designed and built configurable, backward-compatible features for different client networks.
• AI engineering: I designed and built a multi-agent development harness with trace-based evaluation and guardrail hooks, which most of the development team now uses as its standard way of working. To get there, I led team-wide discussions on a shared development process and ran workshops on it.

Modelling Specialist (postdoc) · Eindhoven University of Technology (TU/e) · 12/2023 – 11/2025

Postdoc in an EU consortium project with IBM Research on the modelling and simulation of ReRAM devices.

• Led work package 2 (device modelling and simulation) of a six-work-package EU consortium project with IBM Research, writing its reports and presenting at the project review meetings; coordinated the work of 3 researchers across TU/e and IBM Research and co-supervised a master’s student.
• Built and released LAMMPSKit, an open-source Python package (PyPI, Docker, Read the Docs) with 96% test coverage, used by ~4 researchers at TU/e and IBM Research; refactored it with GitHub Copilot prompt files and skills I built (including a red-green refactoring skill), which 2 other students in my group also used for their own tooling.

Early career fellow (modelling and simulations) · IIT Gandhinagar, India · 09/2022 – 11/2023

A competitively awarded fellowship for my own, independent research project.

• Built a Python package, on top of an open-source graphene toolkit, that generates polycrystalline graphene and hBN structures with graded grain sizes, and used it in molecular-dynamics heat-transport and phonon simulations of graded polycrystalline graphene (up to ~1.04M atoms), writing the Python analysis that identified the competing mechanisms behind its thermal rectification (first author, Carbon 2024).
• Designed and taught, as the sole instructor, a 10-day hands-on graduate course on deep learning for mechanical, molecular and materials engineering, with TensorFlow (Keras) labs in Python, for 22 graduate students from different fields.

PhD candidate (modelling and simulations) · Shanghai Jiao Tong University, China · 09/2018 – 06/2022

• Calibrated an anisotropic (Tsai-Wu) failure criterion for multilayer hBN from over 35 molecular-dynamics failure points, with stress-space accuracy measures I derived: a mean percent offset of −1.43% (SD 5.5%) from the fitted surface, against 23.3% and 25.12% for the Hill and maximum-stress criteria (first author, Mechanics of Materials 2022, front cover).
• Observed stable ripples in sintered hBN by HRTEM and showed with molecular-dynamics simulations that they raise the toughness of multilayer hBN by over 23.4% and its strength by 7.3% at the same time, through distortion-mediated bond strengthening (first author, Acta Materialia 2022).

Education

• PhD, Materials Science and Engineering (Computational Materials Science), Shanghai Jiao Tong University, China, 2018–2022; Outstanding Graduate award
• MSc Physics and BE Mechanical Engineering (dual degree), BITS Pilani, Goa Campus, India, 2013–2018

Licences and certifications

• Databricks Certified Data Engineer Associate (10/2025)
• Microsoft Certified: Azure Fundamentals (AZ-900) (10/2025)

Skills and tools

Python (NumPy, SciPy, JAX, pandas, Pydantic, pytest), SQL, MILP modelling, GitHub Copilot agents, multi-agent orchestration, agent evaluation, Model Context Protocol (MCP), FastAPI, Azure, Databricks, LAMMPS.

Languages

Hindi, Assamese and Bengali: native · English: full professional proficiency (C2) · Dutch: elementary (beginner) · Chinese: elementary (beginner)

Volunteering

Volunteer STEM educator for middle-school students, Shanghai (2018–2020)

Publications, talks, teaching and awards each have their own page: Publications, Talks, Teaching, Awards.