Hi, I’m Jeet Parab, currently working as Technical Lead: Cloud and Data Engineer at HCLTech, with a passion for Machine Learning, LLMs, NLP, and Geospatial Analytics.
I thrive at the intersection of research and engineering, building scalable AI systems that solve real-world problems.
During my research internship at the Sustainability Lab, IIT Gandhinagar, I co-authored SentinelKilnDB (accepted at NeurIPS 2025, Data & Benchmarks Track), contributing to:
- Satellite imagery object detection
- Dataset curation & benchmarking
- Active learning workflows integration
At HCLTech, I focus on deploying enterprise-grade LLMs, optimizing them with RAG and model distillation, and integrating them with cloud platforms for production-ready AI solutions.
Education
B.Tech – Computer Science and Engineering
Indian Institute of Information Technology, Surat Surat, Gujarat | 2021 – 2025
CGPA: 9.03Higher Secondary Certificate (MSBSHSE)
Pace Junior Science College, Thane Thane, Maharashtra | 2019 – 2021
Score: 95.8%Secondary School Certificate (CBSE)
D.A.V. Public School, Thane Thane, Maharashtra | 2007 – 2019
Score: 96.6%
Experience
- Deploying and fine-tuning LLMs in enterprise workflows, advancing AI capabilities with RAG and model distillation.
- Integrating AI solutions with cloud platforms to automate and optimize business processes.
- Delivering scalable, production-ready ML systems with cross-functional collaboration.
- Co-author of SentinelKilnDB (NeurIPS 2025) under Prof. Nipun Batra & Rishabh Mondal.
- Curated a brick kiln detection dataset across South Asia using Sentinel-2 imagery.
- Benchmarked 28 detection models (e.g., DETR, Galileo, TerraMind).
- Improved performance with self-supervised learning (SimCLR, Jigsaw) and active learning.
- Expanded validated detections to 65,616 kilns across South Asia covering India, Nepal, Bangladesh, Afghanistan, and Pakistan.
- Built a multi-functional chatbot using Google Dialogflow & Gemini, reducing workflow time by ~60%.
- Designed and implemented NLP pipelines for enterprise automation.
Projects
LLM-Based Energy Monitoring System – Analytics pipeline to answer natural language energy queries with 90%+ accuracy. Benchmarked five LLMs & optimized prompts.
[Tech: PyTorch, Hugging Face, Streamlit]Signature Verification System – Hybrid ML + OpenCV system achieving precision/recall of 0.92 for user signature verification.
[Tech: Scikit-learn, OpenCV]Crime Monitoring System – REST-based application with API integration for managing and tracking crime reports.
[Tech: React, Node.js, AXIOS]Sentiment-Gender-Bias Analysis – Empirical study on gender bias in LLMs, analyzing sentiment differences across gendered statements.
[Tech: Python, NLTK, Transformers]
Skills
- Python
- C++
- C
- SQL
- PHP
- JavaScript
- HTML / CSS
- PyTorch, TensorFlow, Keras
- Scikit-learn, NLTK
- Hugging Face Transformers
- LLM Fine-tuning & RAG
- Self-Supervised Learning
- Git, Docker, Linux
- Power BI, Figma
- Streamlit, Colab
- Cloud Platforms (AWS, GCP, Azure)
- LLMs & NLP
- Geospatial Analytics
- System Design & Architecture
- Web Development
- Dataset Curation & Benchmarking