Education

PhD Candidate, Artificial Intelligence

University of Amsterdam · ICAI AI for Oncology Lab

Deep representation learning on histopathology images to assess risk of recurrence in breast and ovarian cancers.

MS by Research, Electrical Engineering

Indian Institute of Technology Delhi · GPA 9.0/10

Major in deep learning for computer vision.

BE, Electronics & Communication Engineering

Nitte University, Karnataka · GPA 8.03/10

Major in signal processing, information theory, and digital communication systems.

Experience

PhD Candidate

The Netherlands Cancer Institute

Developing deep-learning models to validate clinical hypotheses in breast and ovarian cancer histopathology, in collaboration with pathologists — quantifying biomarkers and modeling survival and outcome.

AI Advisor (product development)

Diligent Global

Guiding the product team on AI strategy and methodology, from feasibility assessment to incorporation into customer-facing services.

Consultant (computer vision)

Rayreach Technologies

Built deep-learning models to identify personal protective equipment and face masks on factory floors.

R&D Engineer

Skanray Healthcare

Designed hardware and firmware for power delivery, sensory acquisition, and pneumatic actuation of a turbine-based critical care ventilator.

Selected projects

Dundi — AI assistant for Sanskrit scripture study

A domain-specific conversational assistant for the Upanishads, Bhagavad Gita, and Brahma Sutras, with retrieval over verse-keyed commentaries; shipped to production, reaching several hundred active users.

  • NLP
  • LLM applications
  • RAG
  • PostgreSQL
  • Production

AutoComply — Ingredient compliance intelligence

A B2B platform determining cosmetic ingredient status by market (EU CosIng, India CDSCO/BIS, US MoCRA), grounding LLM outputs in official regulation via a RAG layer.

  • LLM workflows
  • Information extraction
  • Next.js
  • Flask/Docker

itsper — Intra-tumoral stroma percentage computer

A command-line utility to compute the tumor-stroma ratio biomarker from tissue segmentations, developed at the AI for Oncology lab.

  • Python
  • PyTorch
  • Open source

Chaos — Lorenz attractor watch face for Watchy (ESP32)

A real-time 3D Lorenz attractor rendered on e-ink — accepted by the Watchy (SQFMI) team and featured in the official gallery. The animation on this site's homepage is the same simulation.

  • C++
  • Embedded systems
  • Real-time graphics