👋 Hi There!
I'M SUKESH S T
Specializing in Generative AI, Agentic AI, and Machine Learning — building production-ready intelligent systems with LLMs, LangGraph, RAG, and modern AI frameworks.
My Projects
Here's what I've been building — from multi-agent AI systems to real-time computer vision
The Problem
Enterprise teams need comprehensive research reports, but manually gathering, analyzing, and synthesizing information from multiple sources is extremely time-consuming and inconsistent.
The Goal
Build an intelligent, multi-agent AI system that automates the entire research pipeline — from planning queries to generating polished, cited reports — with human oversight at critical decision points.
How I Built It
Engineered a stateful, multi-agent pipeline using LangGraph, orchestrating Planner, Researcher, Extractor, Writer, and Reviewer agents. Implemented Human-in-the-Loop (HITL) checkpointing for dynamic re-routing. Developed concurrent async tool-calling across Tavily, Arxiv, and Wikipedia. Deployed on Hugging Face Spaces via FastAPI, SQLite, and Docker.
The Impact
Delivered a fully autonomous research system that generates comprehensive, multi-source reports with source attribution. Human-in-the-loop feedback ensures research quality while reducing manual effort by 80%+.
The Problem
Employees at Flex struggled to quickly find relevant information across hundreds of corporate policy documents, leading to compliance risks and wasted time.
The Goal
Create an AI-powered chatbot that enables natural language querying of corporate policy documents with accurate, source-attributed responses in real time.
How I Built It
Built a RAG pipeline using LangChain, ChromaDB, and SentenceTransformers for semantic search across policy documents. Developed a ChatGPT-style web interface with FastAPI backend, real-time similarity search, and markdown-rendered responses. Deployed as a Dockerized app on Hugging Face Spaces with Groq-hosted LLaMA 3.1 for low-latency inference.
The Impact
Created a production-ready chatbot with sub-second response times, source document attribution, and 95%+ answer relevance. Reduced policy lookup time from minutes to seconds.
Scratch & Dent Detection System
AI-Powered Visual Quality Control
The Problem
At Flex Pvt Ltd, manual visual inspection of manufactured parts for scratches and dents was slow, inconsistent, and costly — leading to quality control bottlenecks on the production line.
The Goal
Develop an automated defect detection system that identifies scratches and dents in real-time using computer vision, reducing dependence on manual inspection.
How I Built It
Built a detection system using YOLOv11, Python, and OpenCV. Trained custom object detection models on labeled defect datasets. Implemented real-time image processing pipelines for continuous quality monitoring on the factory floor.
The Impact
Reduced manual inspection time by 40% via AI-driven quality control. Achieved real-time defect identification with high precision, enabling faster production throughput.
Know Who I Am
A passionate AI/ML Engineer from Chennai, India
I'm Sukesh S T, an AI/ML Engineer specializing in Generative AI, Agentic AI, and Machine Learning. Currently pursuing my B.E. in Computer Science at Vel Tech Engineering College, Chennai with a CGPA of 8.15/10.
I'm passionate about building production-ready intelligent systems using LLMs, LangGraph, RAG, and modern AI frameworks. From designing multi-agent research pipelines to real-time defect detection systems, I thrive on turning cutting-edge AI research into scalable, real-world solutions.
When I'm not coding, I enjoy exploring new AI research papers, contributing to open-source projects, and solving problems on LeetCode.
B.E. Computer Science
Vel Tech Engineering College
2022 — 2026 | CGPA: 8.15ML Intern — Flex Pvt Ltd
YOLOv11 Defect Detection
Jan 20251st Place — Internal SIH
Face Authentication System
2024What I Do
Generative AI & LLMs
Building intelligent applications powered by Large Language Models, prompt engineering, and fine-tuning for domain-specific use cases.
Agentic AI Systems
Designing multi-agent pipelines with LangGraph — coordinating Planner, Researcher, and Writer agents with human-in-the-loop workflows.
Computer Vision & ML
Developing real-time detection and recognition systems using YOLOv11, OpenCV, and classical machine learning algorithms.
Full-Stack AI Applications
End-to-end AI product development with FastAPI backends, modern web frontends, Docker deployments, and cloud infrastructure.
Tech Stack
Languages
AI / ML
Databases
Web & APIs
Tools & Platforms
CRM
My Resume
Education, experience, and qualifications
Career Objective
AI/ML Engineer specializing in Generative AI, Agentic AI, and full-stack AI applications. Experienced in building production-ready intelligent systems using LLMs, LangGraph, RAG, and modern AI frameworks, with a strong focus on scalable, real-world solutions.
Education
Bachelor of Engineering in Computer Science
CGPA: 8.15/10Vel Tech High Tech Dr. Rangarajan Dr. Sakunthala Engineering College, Chennai
2022 — 2026
Relevant Coursework: Data Structures, Algorithms, Machine Learning, DBMS, Networks, OOAD
Higher Secondary Education (HSC)
74.83%SRKBVMHSS, Kulasekharam, Kanniyakumari
2022
Internship
Machine Learning Intern @ Flex Pvt Ltd
6 Jan — 20 Jan 2025
- Built Scratch and Dent Detection System using YOLOv11, Python, and OpenCV
- Reduced manual inspection time by 40% via AI-driven quality control
- Implemented real-time image processing for defect identification
Certifications
Salesforce Platform App Builder
Simplilearn
Aug 2025AWS Certification
Prepsinsta
2025Matlab Onramp
MathWorks
2024