Hi, I'm

Vaishak Girish Kumar

MS in AI at SUNY Buffalo. Building hallucination-resistant VLMs for medical imaging and observability tools for LLM pipelines.

Buffalo, NY

View Research Download Resume Get in Touch

vaishak.sh

$ cat profile.json

{

"name": "Vaishak Girish Kumar",

"role": "AI Researcher & LLM Architect",

"publications": 2

}

$ ls research/

ReXGroundingCT HippoFormer AI-DxMH

$ cat skills

PyTorch • Transformers • VLMs • RAG • LLM

# MICCAI 2025 • Modular LLM architectures

01 / About

About Me

I'm an AI/ML researcher focused on building systems that are both powerful and trustworthy. My current work centers on hallucination-resistant vision-language models for medical imaging and developer tools for understanding RAG pipeline behavior.

I believe the next frontier in AI isn't just better models — it's models we can actually trust and understand. That's why I focus on grounding, observability, and explainability in everything I build.

Previously contributed to open-source NLP infrastructure at Scitonic and published two peer-reviewed papers while completing my B.Tech at Presidency University.

Education

M.S. Engineering Science (AI)

SUNY Buffalo · Expected May 2027

B.Tech Computer Science

Presidency University · 2024

Location

Buffalo, NY

Available for remote collaboration

Publications

Research Projects

OSS Contributions

30%

Perf Improvements

02 / Research

Research & Publications

Focused on making AI systems more reliable and trustworthy, particularly in medical imaging where precision matters most.

ReXGroundingCT

In Progress

Target: ICCAI 2025

View Code

Reference-augmented grounding for vision language models in medical imaging. Combining reference images with query images to improve anatomical grounding and reduce hallucinations in CT analysis.

Implements reference-augmented attention mechanisms
Reduces hallucinations in anatomical predictions by 30%+
Benchmarked on medical imaging datasets

HippoFormer

In Progress

Target: MICCAI 2025

View Code

Transformer-based architecture for hippocampus segmentation from MRI scans. Specialized for medical image segmentation with attention mechanisms.

Transformer-based medical image analysis
State-of-the-art Dice coefficient on benchmark datasets
Efficient inference for clinical deployment

AI-DxMH

Published

IJCRT, Vol. 12, Issue 1

View Code

AI-driven diagnostic framework for healthcare. Fine-tuned LLMs for preliminary diagnostic support in low-resource settings.

92% diagnostic accuracy on target conditions
30% model compression via LoRA for edge deployment
Full-stack web app with <300ms response latency

Peer-Reviewed Publications

AI-DxMH: Artificial Intelligence Diagnosis for Modern Health

First Author

V. G. Kumar, M. F. Pasha, A. Prusty, D. Rajeev, G. Ganesan

International Journal of Creative Research Thoughts (IJCRT) · Vol. 12, Issue 1, January 2024

A comprehensive framework for AI-assisted medical diagnosis in resource-constrained environments, achieving 92% accuracy across multiple conditions.

On-the-fly Prompt Optimization in Multi-Agent Systems: A Comparative Study

Second Author

M. F. Pasha, V. G. Kumar, A. Prusty, S. Taj

International Journal of Creative Research Thoughts (IJCRT) · Vol. 12, Issue 5, May 2024

Evaluating dynamic prompt optimization strategies in multi-agent LLM architectures for improved task performance.

03 / Projects

Featured Projects

A selection of research projects and tools focused on making AI systems more reliable, observable, and trustworthy.

ReXGroundingCT

Reference-augmented grounding for vision language models in medical imaging. Reduces hallucinations in anatomical predictions using foveated attention mechanisms.

PyTorchVision TransformersMedical AI

VectorLens

RAG observability tool for detecting and debugging retrieval pipeline failures. Provides visual insights into vector search quality and context relevance.

PythonFastAPIChromaDBReact

AgentReplay

Time-travel debugging framework for LLM agent workflows with DAG-based replay, visual context diffs, and execution tracing.

TypeScriptReactLangChain

HippoFormer

Transformer architecture for hippocampus segmentation from MRI scans. Achieves state-of-the-art Dice scores on medical imaging benchmarks.

PyTorchHuggingFaceMedical Imaging

Agentic RAG Pipeline

Multi-agent retrieval system with self-correction, query decomposition, and iterative refinement for complex document Q&A.

LangGraphPythonRAG

AI-DxMH

AI-driven diagnostic framework for healthcare. Fine-tuned LLMs achieving 92% accuracy with 30% model compression via LoRA.

PyTorchTransformersHealthcare

LLM Eval Suite

Comprehensive evaluation framework for testing LLM reliability, hallucination rates, and factual accuracy across domains.

PythonPytestOpenAI

Medical VLM Benchmark

Standardized benchmark suite for evaluating vision-language models on medical imaging tasks with clinical ground truth.

PyTorchMedical AIBenchmarking

View All on GitHub

04 / Experience

Experience & Skills

Open-Source Developer

Scitonic via Tonic.AI · Remote

Jan 2024 – Jul 2024

Core contributor to LLM pipeline infrastructure through Tonic.AI's open-source program, focusing on NLP tooling and data processing optimization.

Fixed 17 critical bugs across tokenization, preprocessing, and API modules

Refactored 1,500+ lines of Python code for improved maintainability

Achieved 30% performance improvement in pandas preprocessing pipelines

Authored regression tests for 10+ production modules

Technical Skills

Languages

PythonSQLTypeScriptLaTeXBash

ML / Deep Learning

PyTorchHuggingFaceScikit-learnOpenCVNumPy

LLM / NLP

Fine-TuningRAG SystemsLangGraphVision-Language ModelsPrompt Engineering

Infrastructure

FastAPIPostgreSQLDockerNext.jsGit

05 / Contact

Get in Touch

Interested in collaboration, research opportunities, or discussing AI? I'm always open to connecting with fellow researchers and engineers.

vaishakg@buffalo.edu

Best for research inquiries

GitHub

gustav-proxi

Check out my code

vaishakgkumar

Let's connect

Location

Buffalo, NY

Eastern Time Zone