Sitemap
A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.
Pages
Vijay Vankadaru
ML Research Engineer
Posts
Hallucination is Structural, Not Accidental
Published:
The standard framing of LLM hallucination is as a bug. Train better. Prompt better. Retrieve better. The bug will eventually go away.
Why Multi-Instance Learning is Actually Beautiful
Published:
There’s a class of problems in ML that most supervised learning frameworks can’t handle cleanly: you know the label for a group of examples, but not for any individual one.
CTC Alignment and Why Temporal Correspondence Matters in Multimodal Learning
Published:
When you fuse audio and text representations, the obvious approach is to encode both independently and then concatenate or cross-attend. It works. But it misses something important: the correspondence between what was said and how it was said, at the same moment in time.
What It Felt Like to Finish a Paper
Published:
The hallucination survey went live on MetaArXiv in March. I want to write about what the process was actually like before the memory fades.
Industry Experience Is a Research Asset, Not a Gap to Apologize For
Published:
There’s a version of the story I told about myself for a while that went like this: I spent years doing applied work in industry before getting serious about research. That framing treated the industry work as a detour — something to acknowledge and move past.
Learning to Actually Read Papers
Published:
Nobody teaches you how to read a paper. You’re expected to figure it out, and most people do eventually, but the path is inefficient and kind of humbling.
What Research Meetings Actually Feel Like
Published:
I had my first real research meeting with Prof. Paulik in late August. Not a class, not office hours — a working meeting about a project I was contributing to. I want to write down what it felt like before I forget.
The Gap Between Building ML Systems and Doing ML Research
Published:
I started at DASION in 2021 as a high school intern. By the time I enrolled at Berkeley this month, I had spent three years building ML systems that actually ran in clinical settings — models that processed real patient data, infrastructure that stayed up at 99.9%, pipelines that clinicians depended on. I thought that experience would translate directly to research.
portfolio
Rethinking Medical LLM Hallucinations: A System-Level Survey
Co-authored survey arguing hallucination in medical LLMs is a structural property of probabilistic generation. Published MetaArXiv March 2026.
Multimodal Multi-Instance Learning for Depression Detection
First multimodal MIL framework combining Wav2Vec 2.0 audio with MT5/RoBERTa text via CTC temporal alignment for depression detection on DAIC-WOZ.
PedRAG: Retrieval-Augmented Generation for Pediatric Medical QA
RAG pipeline for pediatric medical question-answering combining dense and sparse retrieval with age-specific classification. 34% accuracy improvement, 42% hallucination reduction over baseline.
AGMNT — E-commerce Recommendation System
Production recommendation engine and full-stack platform connecting Asian brands to Western markets, serving 10K+ users across 15+ brand partners.
publications
Multimodal Multi-Instance Learning for Depression Detection: Combining Wav2Vec 2.0 Audio and Transformer Text Features with CTC Temporal Alignment
Published in Target: NeurIPS 2026, 2026
First multimodal MIL framework for depression detection on DAIC-WOZ. Combines MT5/RoBERTa text with Wav2Vec 2.0 audio via CTC temporal alignment. Achieves F1>0.90, surpassing prior SOTA. Directly addresses interviewer bias via strict prompt exclusion.
Recommended citation: Vankadaru, V. et al. (2026). Multimodal Multi-Instance Learning for Depression Detection. Target: NeurIPS 2026.
PedRAG: Retrieval-Augmented Generation for Pediatric Medical Question-Answering
Published in Target: ICML 2026 (Poster), 2026
RAG framework for pediatric medical QA using dual-retrieval architecture combining dense encoders and sparse retrieval (BM25) with age-specific classification. Achieves 34% accuracy improvement and 42% hallucination reduction over baselines.
Recommended citation: Vankadaru, V. et al. (2026). PedRAG: Retrieval-Augmented Generation for Pediatric Medical QA. Target: ICML 2026.
Rethinking Medical LLM Hallucinations: A System-Level Survey
Published in MetaArXiv, 2026
A systems-level survey arguing that hallucination in medical LLMs is a structural property of probabilistic generation, not a fixable bug. Synthesizes 50+ papers on detection, mitigation, and benchmarking through a risk management lens.
Recommended citation: Matthews, A., Vankadaru, V., Roosta, T., & Passban, P. (2026). Rethinking Medical LLM Hallucinations: A System-Level Survey. MetaArXiv.
Download Paper
talks
Talk 1 on Relevant Topic in Your Field
Published:
This is a description of your talk, which is a markdown file that can be all markdown-ified like any other post. Yay markdown!
teaching
Teaching experience 2
Workshop, University 1, Department, 2015
This is a description of a teaching experience. You can use markdown like any other post.