About
An AI engineer from Nepal who splits his time between shipping edtech systems and doing medical AI research.
My Story

I build AI systems and write up what I learn. Most of it sits around agents, retrieval, healthcare ML and knowledge graphs. I keep ending up in places where a wrong answer costs something — a clinic, a classroom, a pipeline that feeds another pipeline — so I care a lot about evaluation.
Since July 2026 I'm an AI Engineer at Amrita AI, working on Ample LMS, our other edtech systems, and the research behind them. I joined as an Elite AI Research Assistant in December 2025. Alongside that, I'm a Research Associate in Medical AI with Harvard Medical School & Dana-Farber Cancer Institute. My four papers are in Elsevier's Procedia Computer Science and Springer conference proceedings (ICTCS).
In 2024 I won the NASA Space Apps Challenge Local Impact Award with a system we built in 48 hours. I finished my B.Tech in Computer Science (AI) at Amrita in 2026, with my first paper already out from second year. If there's a thread, it's this: build it, test it properly, write down what went wrong.
Journey
AI Engineer
Amrita AI (Amrita School of AI)
Engineering Ample LMS and other edtech systems, and carrying the research behind them — retrieval, agentic tutoring and evaluation — into production.
Elite AI Research Assistant
Amrita AI (Amrita School of AI, Faridabad)
Built machine learning models and tooling for medical AI research. Presented Medsum at the India AI Impact Summit 2026.
Research Associate – Medical AI
Harvard Medical School & Dana-Farber Cancer Institute
Developing medical AI models for histopathology and clinical signal analysis, with scalable data pipelines and interpretable tools supporting real-time clinical use.
QA Engineer Intern
Heapvue
Built and maintained automated test scripts using Selenium Python, improving testing efficiency by 40%. Performed API testing with Postman and collaborated with developers to ensure quality across microservices.
AI Research Intern
Amrita Vishwa Vidyapeetham
Developed and researched prostate cancer grading systems using deep learning.
B.Tech Computer Science (AI)
Amrita Vishwa Vidyapeetham, Amritapuri
Graduated in 2026 with a specialization in AI and a CGPA of 9.18/10. Published my first conference paper as an undergraduate (ICTCS 2023, Springer).
Values
Research as Practice
Publish findings, share methodology, and treat every project as an opportunity to contribute something reproducible to the field.
Build What Ships
A system that runs on your laptop is a prototype. A system that handles edge cases, has monitoring, and doesn't crash at 2 AM is engineering.
Write to Think
Writing clarifies ideas that feel clear in your head. Technical writing, papers, and post-mortems are how I refine what I think I know.
Compound Over Time
Skills, publications, and open-source contributions accumulate. The goal isn't a single big break. It's consistency across years.
Bio
Research Interests
Frequently Asked Questions
Who is Avishek Rauniyar?
Avishek Rauniyar is an AI Engineer from Nepal. He works at Amrita AI on Ample LMS and edtech systems, and does medical AI research with Harvard Medical School and Dana-Farber Cancer Institute. He has four publications and won the NASA Space Apps Challenge Local Impact Award in 2024.
What does Avishek Rauniyar research?
His research areas include agentic AI systems, healthcare AI, knowledge graphs, retrieval-augmented generation (RAG), multi-agent architectures, and clinical decision support systems. He has published in Springer and Elsevier venues.
Where does Avishek Rauniyar work?
He is an AI Engineer at Amrita AI (Amrita School of AI) since July 2026, after joining as an Elite AI Research Assistant in December 2025. He also works as a Research Associate in Medical AI with Harvard Medical School & Dana-Farber Cancer Institute. He is originally from Barahakshetra, Nepal.
What projects has Avishek Rauniyar built?
Besides Ample LMS at work, his projects include CogniMail AI (multi-agent email intelligence), Omkara Extractor (knowledge graph construction pipeline), GraphRAG Movie Recommendation Engine, and SageLoRa (parameter-efficient LLM fine-tuning). See the projects page for detailed case studies.