Sujit Laware
I’m a final-year dual-degree student from IIT Madras and SRMIST, building at the intersection of AI and backend engineering. I’ve shipped production systems across three internships — from fine-tuning LLMs and building MLOps pipelines to leading AI product development at an IIT Madras startup.
Outside work, I’ve won hackathons, led a 1,000+ member AWS community, and contributed to a patent at IIT Madras.
Experience
Research Intern
RuTAG (Rural Technology Action Group) - IIT Madras
Chennai, India
Built and optimized deep learning models (ResNet-18, ViT) on multi-class agricultural image data, later extending to regression for continuous moisture estimation, achieving 89% accuracy in drying-stage classification for cardamom.
Integrated image-based predictions with load-cell weight data via a Raspberry Pi–based controller and real-time dashboard, and contributed to an end-to-end image-to-weaving automation system using OpenCV and CNNs, including work toward a filed provisional patent for an AI-driven heat-pump dryer system.
Technical Lead (AI and Cloud)
Zeex AI (IIT Madras Pre-Incubated Student Startup)
Chennai, India
Led the development of Z-Tracs, a traffic management system recognized in the final round of the CUMTA (Chennai Unified Metropolitan Transport Authority) Open Innovation Challenge. Led the creation of Z-Market, an AI-powered security analytics solution tailored for supermarkets, transforming surveillance systems into intelligence platforms.
AI/ML and Cloud Engineering Intern
Run Verve
Chennai, India
Led the backend and deployment intern team for the RunVerve application by building scalable microservices, developing and testing REST APIs using FastAPI and PostgreSQL, and managing AWS cloud infrastructure, Docker, and CI/CD workflows for real-time production deployment.
Developed ML models for fitness analytics and worked on VDocs, an NER-based system that extracts medicine, dosage, strength, and timing from prescriptions.
Education
Indian Institute of Technology, Madras
B.S.
GPA: 7.00
SRM Institute of Science and Technology
B.Tech.
GPA: 8.40
Skills
Projects
AI Tutor
Fine-tuned Mistral-7B using QLoRA (PEFT) on 800 instruction samples, reducing training loss from 1.18 → 0.52 on a single Colab T4 GPU. Enabled memory-efficient training via 4-bit NF4 quantization, LoRA adapters, and gradient checkpointing.
Augmented inference with a RAG pipeline using ChromaDB and sentence-transformers for context-aware responses; evaluated on 30 prompts with 70%+ improvement in response clarity vs base model.
Hyperlocal Grocery Delivery Platform
Built a full-stack Zepto-style delivery platform with React frontend and FastAPI backend; Kafka-driven order state machine, Redis inventory cache for sub-10ms reads, and PostGIS nearest-store routing. Integrated Claude API for smart reorder suggestions and XGBoost demand forecasting; deployed on AWS via Docker with CI/CD.
US Visa Approval Prediction
Built an end-to-end MLOps pipeline to predict U.S. visa approvals using applicant demographic and employment data. Automated data ingestion and preprocessing with Apache Airflow; tracked experiments and model versions via MLflow. Containerized with Docker and deployed on AWS with CI/CD integration. Achieved 87% accuracy with continuous monitoring and automated retraining.
Honors and Awards
Finalist, Google Agentic AI Hackathon 2025
Agentic AI financial copilot
Winner, Block-a-thon – AI & Blockchain Hackathon
AI agent–based DeFi yield optimization protocol
Runner-up, Avishkar Hackathon (NIOT & SRMIST)
Digital twin for predictive pipeline maintenance
Leadership & Community
AWS Student Builder Group – IIT Madras
Led a 1,000+ member AWS cloud community by organizing and hosting 7 technical events, including speaker sessions and hands-on workshops to help students learn AWS and cloud computing.
Paradox, TechXConf 2024 & GDG Cloud Community Day
Coordinated sponsorships for Paradox at IIT Madras and contributed as event crew member at national-level AI and cloud conferences including TechXConf 2024 and Google Developer Groups (Chennai) Cloud Community Day.