Portfolio - Vishal Krishna Kumar

I develop intelligent ML systems with a strong focus on accuracy, explainability, and robust evaluation, ensuring reliable performance through rigorous experimentation.

This is me.

Hi, I'm Vishal 👋

I'm pursuing an MS in Computer Science (Data Science concentration) at City University of Seattle (Expected 2027), where I'm a Dean's Scholar and a graduate researcher. I hold a B.Tech in Computer Science and Engineering from SRM Institute of Science and Technology. I enjoy translating complex problems into practical, well-designed technical solutions with real-world impact.

Currently, my interests are centered around Generative AI, Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and data-driven system design. I recently completed an AI Engineer internship at Hirello.ai (July – September 2026), and I'm looking for full-time software engineering roles starting in 2027.

My Stack

ml_ai

Machine Learning

Deep Learning

NLP

Computer Vision

RAG

GraphRAG

frameworks

PyTorch

TensorFlow

Scikit-learn

OpenCV

Hugging Face

LangChain

database

MySQL

Pinecone / FAISS

PostgreSQL

tools

Git

Docker

AWS

Neo4j

My Experience

Hirello.ai — Remote

AI Engineer Intern

Built a resume-grounded RAG application-autofill feature for job-portal forms. Contributed code to 5 of the platform's 10 agents (Resume Optimizer, LinkedIn Labs, Outreach, Action Academy, Chat with Hiro). Fixed 3 Resume Optimizer defects, repaired Gmail, Outlook and Yahoo OAuth connection problems, fixed an IDOR access-control vulnerability with regression tests, and added cross-browser end-to-end tests covering 3 browsers. Worked in a 7-person product and engineering team across 10 one-week sprints. Stack: Python, FastAPI, SQLAlchemy, MySQL, JavaScript, Vue 3, Vertex AI/Gemini, Pinecone.

July 2026 – September 2026

City University of Seattle

Graduate Researcher — Applied ML and Intelligent Systems

Developed a modular RAG pipeline using dense and sparse retrieval, structured chunking and graph reasoning, reducing LLM hallucinations by 30%. Built reproducible experiment workflows, reducing debugging and rerun time by 40%. Worked with PyTorch computer-vision and NLP models.

August 2025 – Present

Illinois Institute of Technology

Graduate Teaching Assistant (part-time) — Computer Networks

Supported networking labs and graded assignments. Assisted students using the FABRIC testbed. Helped configure and troubleshoot multiple servers.

January 2025 – May 2025

Job Simulations (Forage) — not employment

BCG (via Forage)

Job simulation (Forage) — Data Science

Customer churn prediction: random forest on 10,000+ records; recall improved from 32% to 50%.

Walmart (via Forage)

Job simulation (Forage) — Advanced Software Engineering

Implemented a heap-based priority queue in Java for shipping workflows, produced a UML class diagram for a data processor with multiple execution modes and database connectivity, and designed an ER diagram for a new accounting database based on business requirements.