Looking for a job

CHIRAG GARG

@chiraggarg1

Message

I build reliable LLM, RAG, climate ML, and production AI systems with Python and PyTorch.

India

Message

What I'm looking for

I'm looking to build reliable, scalable ML and LLM systems, especially across retrieval, evaluation, climate ML, production backend services, and on-device inference.

At GIST Impact, I build production climate-risk and AI systems, including a RAG hallucination-detection framework that achieved approximately 88% accuracy and reduced LLM context by approximately 20%.

I develop config-driven climate pipelines, bias-correction models, FastAPI services, and geospatial risk-assessment modules using AWS S3, MySQL, PostgreSQL, and PostGIS. My work spans CMIP6-to-IMD climate mapping, retrieval optimization, model evaluation, and scalable spatial analysis.

I also build practical LLM infrastructure and local AI products: a PyPI developer toolkit, an agentic resume-tailoring system, and FileOS, an on-device semantic file manager using hybrid retrieval, SigLIP, and ONNX Runtime.

Experience

Work history, roles, and key accomplishments

Current

ML Learning

Current

Personal Project

Aug 2026 - Present (1 month)

Developing ML and deep-learning foundations in PyTorch, covering data preparation with Polars and Albumentations, NLP preprocessing, embeddings, training, evaluation, and MLflow experiment tracking. Implementing character-based embeddings and Transformer architectures from first principles.

PyTorch Polars Albumentations NLP Embeddings Transformers Multi Head Attention MLFlow

Current

AI/ML Engineer

Current

GIST Impact

May 2025 - Present (1 year 4 months)

Built a RAG hallucination detection framework using embedding similarity, achieving ~88% accuracy and reducing LLM context by ~20%. Developed production-ready climate-risk pipelines and risk-assessment modules with FastAPI, MySQL/PostgreSQL, and PostGIS.

Job Application Assistant

Personal Project

Apr 2026 - Jul 2026 (3 months)

Built an end-to-end RAG system with ChromaDB, sentence-transformer embeddings, and fusion reranking to tailor resume content to job descriptions. Engineered a stateful LangGraph workflow with human review, parallel LLM generation, SQLite checkpointing, SSE streaming, and resume scoring.

RAG ChromaDB Sentence Transformers Reranking LangGraph SQLite SSE Resume Scoring LLM Usage Management

FileOS

Personal Project

Jan 2026 - Mar 2026 (2 months)

Built a local-first semantic and hybrid retrieval system combining dense embeddings, BM25, vector search, and similarity graphs for intelligent file discovery. Added multimodal and on-device ML using SigLIP and ONNX Runtime for image understanding and zero-shot classification.

Dense Embeddings BM25 Vector Search Similarity Graphs SigLIP ONNX Runtime Zero Shot Classification

Python Developer & LLM Infrastructure Toolkit

Infrakit

Oct 2025 - Dec 2025 (2 months)

Published a reusable Python developer toolkit on PyPI providing configuration management, structured logging, session tracing, log rotation, project scaffolding, and dependency analysis. Built a multi-provider LLM client with API-key rotation, persistent quota tracking, structured outputs, and concurrent/async inference.