Open to opportunities

NIRMAL KUMAR

@nirmalkumar1

Message

AI/ML engineer building production-grade CV and NLP systems to deliver measurable business value.

India

Message

What I'm looking for

I’m looking for an AI/ML role where I can build end-to-end, production-grade CV and NLP systems—RAG, LLM orchestration, and MLOps on AWS/Azure—working with a team that values measurable impact, clean engineering, and continuous improvement.

I’m an AI/ML Engineer with 3 years of experience building and deploying production-grade AI solutions across insurance tech, industrial automation, safety, and customer engagement. I focus on end-to-end AI system development—from Computer Vision (YOLO, OCR, depth estimation) to advanced NLP (RAG architectures, LLM orchestration, conversational AI)—to solve complex business problems.

In insurance tech, I built a RAG-powered document intelligence system for health insurance policies, including multi-namespace vector indexing and OCR pipelines for scanned PDFs. I also developed an insurance WhatsApp chatbot with intent classification routing and channel-specific session management, plus a Fireflies Attribution Agent that automates marketing attribution from sales call transcripts.

I’ve delivered real-world computer vision systems like a YOLOv8-based ANPR pipeline for Indian truck license plates, PPE compliance monitoring with YOLOv11 and automated notifications, and volumetric truck load estimation using YOLOv8 + ZoeDepth point cloud analysis. I care about scalable deployment and reliability—using FastAPI microservices, Pinecone/FAISS, Docker, and CI/CD on AWS/Azure to keep systems measurable, maintainable, and ready for production.

Experience

Work history, roles, and key accomplishments

AI/ML Engineer

Beshak.org

Oct 2025 - Mar 2026 (5 months)

Built a RAG-powered health insurance policy intelligence system to extract and query documents using Pinecone, PostgreSQL, and Tesseract OCR. Developed an insurance WhatsApp advisory chatbot with intent routing and channel-specific session management, plus a Fireflies attribution agent with scheduled ETL and automated reporting.

fastAPI RAG Pinecone PostgreSQL Tesseract Webhooks

AI/ML Engineer

TFrenzy

Jun 2025 - Sep 2025 (3 months)

Implemented a RAG assistant using AutoGen with custom MCP servers (LinkedIn/website/web search), Pinecone retrieval, Azure DevOps CI/CD, and Docker for scalable Azure Container Apps deployment. Developed an end-to-end ANPR system (custom YOLOv8 + PaddleOCR), truck load volumetric estimation using YOLOv8 and ZoeDepth, and a real-time PPE compliance monitoring workflow with automated alerts.

AutoGen MCP Pinecone Azure DevOps Docker YOLOv8 PaddleOCR Streamlit fastAPI ZoeDepth Paddle OCR

Data Scientist / MLE

UCT Consultancy

May 2024 - Jun 2025 (1 year 1 month)

Built a real-time helmet detection system for workplace safety using YOLOv8 with automated alerts and achieved 91.7% mAP@50. Developed multiple AI applications including a legal document query chatbot with GPT-3.5 Turbo + RAG, an AI hotel booking assistant using LangChain/LangGraph and vector databases, an interview simulator with Whisper ASR and YOLO/MediaPipe proctoring, and a sales forecasting

YOLOv8 RAG fastAPI LangGraph MediaPipe Face Mesh SARIMA Prophet DVC Lang Chain

Machine Learning Engineer

Refactor Private Limited

May 2023 - Apr 2024 (11 months)

Developed a Power BI credit card financial dashboard for real-time fraud detection and risk monitoring using SQL transaction data. Built customer segmentation and classification using K-Means and Logistic Regression, deployed with Docker and AWS S3 using CI/CD, and created a 98.6%-accuracy VGG16-based kidney disease classifier with MLflow/DVC tracking and AWS deployment via Flask API.