Tanay Kinariwala

Computer Engineering Student | AI/ML Research Sardar Patel Institute of Technology
Mumbai, India

Tanay Kinariwala

I am a fourth-year Computer Engineering student. My research interests lie at the intersection of large language models, retrieval-augmented generation, multi-agent systems, NLP, and applied machine learning. I am particularly interested in building reliable, knowledge-grounded AI systems that can reason over complex information and operate in real-world environments.

Currently seeking research collaborations and internship opportunities in AI/ML, particularly involving agentic systems and knowledge-enhanced AI.

Education

B.Tech. in Computer Engineering

Sardar Patel Institute of Technology, Mumbai | 2023 – 2027 | CGPA: 9.53/10

Minor in Management

SP Jain Institute of Management and Research | 2024 – 2026 | CGPA: 8.50/10

Publications

A Sovereignty-Preserving Enterprise AI Copilot with Hybrid Retrieval and Self-Healing Text-to-SQL

Accepted · ICSL-DSGA 2026

Proposed a model-agnostic Enterprise AI framework integrating Hybrid RAG, self-healing Text-to-SQL, and RBAC to preserve data sovereignty.

Multi-Agent RAG for Automated Grading

Accepted · 18th International Conference on Contemporary Computing (IC3)

Developed a Multi-Agent RAG framework combining Actor–Critic Reflexion achieving 98.97% human correlation while minimizing hallucinations.

Two-Stage Time-Topological Optimization for Demand-Responsive Transit Design

Accepted · 12th CTSEM

Created a genetic optimization framework using 2.1B+ GPS records that reduced vehicle kilometers traveled by 47.6%.

Machine Learning–Driven Weekly Momentum Stock Selection System

Presented · 8th Pan-IIT International Conference on Financial Markets [Certificate]

Proposed a momentum investing framework using walk-forward validation across 4,000+ stocks yielding a +5.59% weekly alpha.

Cryptographic API Misuse Detection Using BERT-Based Model

Published · ICITIIT 2026 (IEEE Xplore) [IEEE]

Developed a BERT-based deep learning framework capable of automating the identification of cryptographic API misuses directly from source code.

Experience

Research Collaborator, University of Illinois Urbana-Champaign (UIUC)

Jun 2026 – Present | Remote

Built LightGBM models predicting checkpoint failures across 104K+ SWE-agent runs, achieving 0.871 ROC-AUC.

Research Intern, IIT Bombay

Jan 2026 – Jun 2026 | AI & Intelligent Transportation Systems

Developed a scalable geospatial analytics pipeline and optimization framework, reducing transit route cycle times by 49.6%.

Software Engineer Intern, Deloitte

Jun 2025 – Aug 2025 | Technology & Transformation

Developed and deployed a production-ready, full-stack web application for a major banking client.

Freelance Software Developer

May 2025 – Present

Led the end-to-end development and deployment of three applications for businesses across the healthcare and manufacturing domains.

Selected Projects

Enterprise AI Platform

React, FastAPI, PostgreSQL, LangChain, LLMs

Built a secure, RBAC-enabled platform supporting RAG-based natural language interaction with PDFs, CSVs, and SQL databases.

Automated Answer Sheet Evaluation System

Python, OCR, RAG, LLM APIs

Developed an evaluation pipeline integrating OCR and RAG to automate rubric-aware grading across 1,000+ descriptive answer sheets.

Momentum Investing using Machine Learning

Python, XGBoost, Scikit-learn, Pandas

Engineered an end-to-end ML framework for momentum stock selection backtested over 60+ weeks of historical market data.

Technical Skills

Programming: Python, Java, C, JavaScript, SQL
AI / ML: PyTorch, TensorFlow, Scikit-learn, LangChain, Hugging Face, OpenCV, XGBoost
Web / Cloud: React.js, Next.js, Node.js, FastAPI, PostgreSQL, MongoDB, Docker, AWS
Others: RAG, LLM Agents, Multi-Agent Systems, NLP, Vector Databases (FAISS, ChromaDB)

Leadership

Vice Chairperson

Sardar Patel Consulting Group | Aug 2025 – Present

Managed end-to-end client engagement for 4+ industry clients, delivering technology consulting solutions.

Joint Secretary

Rotaract Club of SPIT | Aug 2024 – Jul 2025

Coordinated 100+ members and executed 20+ large-scale events with cumulative participation exceeding 4,000 attendees.

Director General

Model United Nations, SPIT | Jan 2025

Directed and organized a university-level MUN conference overseeing planning and cross-functional execution for 500+ participants.

Achievements

Winner — HackNova, 24-hour Hackathon (125+ teams)
Winner — SPIT Oddesys Frontend Hackathon (50+ teams)
Winner — SPIT BCS Business Case Study (60+ teams)