About kunal
I am a Computer Science graduate student at Rutgers (MS, GPA 3.83), and I build reliable, scalable systems end to end. Depending on the problem, that looks like backend and distributed services, LLM and multi-agent applications, or data and streaming pipelines. What ties it all together is caring about how systems behave in production: performance, reliability, and clean engineering.A few things I have built: Synapse, a multi-agent LLM platform that scores 76% on GAIA Level 1, which is 18 points above GPT-4o; QueryPilot, a natural-language-to-SQL agent that reaches 81% on the Spider benchmark with a QLoRA fine-tuned model; and PulseQueue, a distributed event-streaming and observability platform built with Java, Spring Boot, Kafka Streams, and AWS. I am also a co-author on CultSportQA, an EMNLP 2025 paper introducing a multilingual QA benchmark across 60 countries and 11 languages. Right now I am an AI Automation Intern at National Appraisal Partners, where I built an end-to-end ML system that replaced a fully manual pricing workflow.Day-to-day, I work across Python, Java, FastAPI, Spring Boot, LangGraph, PyTorch, Kafka, PostgreSQL, Docker, and AWS. I am open to Software Engineer, Machine Learning, and Data roles where I can build systems end-to-end, so feel free to reach out at [email protected].
Key Skills
Experience
Ai Automation Intern
CurrentNational Appraisal Partners
* Built an end-to-end ML system that predicts real estate appraisal fees and recommends appraisers from raw bid inputs, replacing a fully manual pricing workflow * Trained CatBoost regression and classification models that improved fee prediction from an R² of 0.02 to 0.39 (MAE around $350) through feature engineering across 364 input fields * Engineered an ETL pipeline that normalizes 8,000+ historical bids from FileMaker Pro, with shared preprocessing that guarantees exact train and serve parity * Deployed the models as a Flask REST API integrated into Google Sheets through Apps Script, giving the bidding team one-click predictions inside their existing workflow
Parttime Lecturer
Rutgers University–new Brunswick
* Taught introductory data science using R, covering data manipulation, visualization, and statistical analysis * Guided students through hands-on coding exercises and real-world datasets to build foundational programming and analytical skills
Parttime Lecturer
Rutgers University–new Brunswick
* Taught Discrete Structures including probability, statistics, and graph theory, covering core mathematical foundations for computer science * Led problem-solving sessions and explained algorithmic applications of discrete concepts to improve student understanding
Education
Rutgers University–new Brunswick
Master Of Science
Bhartiya Vidya Bhavans Sardar Patel Institute Of Technology Munshi Nagar Andheri Mumbai
Bachelor Of Technology
Mvm College Of Commerce And Science
Interested in connecting with kunal?
Sign up for NinjaHire to send a connection request.
Common Questions
What is kunal's expertise?
kunal specializes in LLM Engineer, with expertise in a/b testing, ai agents, apache kafka, c++, data engineering.
Where is kunal located?
kunal is based in new brunswick, new jersey, united states.
How much experience does kunal have?
kunal has 3+ years of professional experience.
How can I contact kunal?
You can connect with kunal through NinjaHire by signing up for a free account.
Other LLM Engineers
Looking for a different LLM Engineer?
Describe exactly who you need and NinjaHire will source them for you.
Type a role to try NinjaHire for free