Software & Machine Learning Engineer — Research prototypes, dependable developer systems.
New Delhi, India.
I build the layer where ambitious research becomes useful software. I work across model behavior, APIs, deployment, and the product decisions that make intelligent systems useful.
Student Council Web Ecosystem
Pretraining and Benchmarking Small Language Models
Layer5 documentation and Meshery.io
Two editable entries for research and data systems work.
Groundwork on SICKLE++
Real-time data integration and interactive visualization
A multimodal sarcasm detection tool combining image and text analysis. It combines BART encoders with a ViT backbone and extracts image segments, based on target-augmented shared fusion research.
A creator-first social platform shaped as a set of services: content monetization, real-time chat, media sharing, and an NFT marketplace with Razorpay and blockchain-backed transaction logging.
A study ecospace and agentic study assistant. Python and FastAPI power asynchronous endpoints for personalized quizzes and study content through pluggable LLM adapters.
An extension of the Quench benchmark for reasoning in Indic contexts, adding bias injection, three languages, and structured cause-effect generation in Boolean logic for reproducible evaluation.
Python, JavaScript, TypeScript, Go, Java, C++, R
PyTorch, TensorFlow, LLMs, model serving, inference optimization
FastAPI, Flask, Express, REST APIs, async services
PostgreSQL, MongoDB, Redis, MySQL, Elasticsearch
Docker, Kubernetes, GitHub Actions, Git, GitLab
React, Next.js, responsive interfaces, accessibility
Placeholder record for a publication or working paper. Add the final title, co-authors, venue, year, abstract, and DOI when the owner is ready to publish the work.
For thoughtful collaborations, difficult systems questions, or a good research lead, email is the most direct route.