AI / ML ENGINEER & SOFTWARE DEVELOPER
From computer vision and local-first AI to autonomous drone systems, I build practical software at the point where research meets real use.
Learning by building,
testing, and refining.
Projects across intelligent agriculture, document automation, robotics, and decision support. Each one started with a practical problem.
A livestock identification platform that recognises 40+ Indian cattle breeds from a smartphone camera, paired with a multilingual app and a retrieval-augmented chatbot.
Built around a fine-tuned deep-learning model, Firebase real-time data, and a FastAPI backend designed with rural connectivity in mind.
An RL stabilisation framework for multirotor drones navigating without GPS, trained with PPO for more stable flight under uncertain conditions.
Trained for 600k+ timesteps in Unreal Engine 5 and AirSim to suppress rotational instability and INS drift under stochastic wind disturbances.
A command-centre concept for detecting supply-chain disruptions, identifying affected routes, and simulating recovery options.
Designed to help turn risk signals into a fast, safe response by combining document and text analysis with route-level decision support.
A precision-agriculture dashboard for drone-assisted farm monitoring, combining environmental data, crop-health analytics, and drone navigation in one interface.
Includes NDVI-oriented crop stress visualisation, mapping and field coverage, environmental monitoring, and an AI farming assistant.
A local-first Windows desktop application that extracts purchase-order fields with PDF/DOCX parsing, Tesseract OCR, pattern matching, and an on-device Ollama LLM.
Generates templated POs, MoUs, trainer invoices, offer letters, and internship certificates. Includes trainer, intern, company, and finance modules with SMTP email and a packaged installer.
Additional product and analysis work beyond the five projects listed on my résumé.
A pharmacy management MVP focused on inventory visibility, expiry monitoring, email alerts, and role-based operations for managers and administrators.
Built as a fast, responsive web application with a practical focus on day-to-day pharmacy workflows.
An additional analysis project maintained in a separate repository.
The repository is currently private and has limited public documentation, so I have not inferred additional features or results.
* Core project descriptions and stacks follow my résumé. Repository links are included where I could confirm a public project; internal, private, or undocumented work is labelled accordingly.
Internships and research that shaped how I approach software, intelligent systems, and engineering trade-offs.
Built DocGen, a local-first Windows desktop application automating internal document workflows. Implemented PDF/DOCX parsing, OCR, pattern-based extraction and on-device LLM field extraction; developed templated document generation for POs, MoUs, invoices, offer letters and certificates. Built trainer, intern, MoU-company and finance modules with SMTP email, then packaged a Windows installer.
Developing reinforcement-learning policies for autonomous drone navigation in GPS-denied environments using Unreal Engine 5 and Cosys-AirSim. Designed reward functions, state representations, and evaluation metrics; optimised simulation runtime and environment interaction pipelines.
Built a drone-based precision-agriculture dashboard with crop-health analytics, NDVI stress detection, and an integrated AI farming-assistant chatbot.
A practical engineering mindset, a researcher's curiosity, and a preference for building end-to-end systems that can actually be used.
Kattankulathur (KTR), Chennai
B.Tech · Computer Science and Engineering (AI & ML)
I enjoy connecting the full pipeline: data and model experiments, backend services, product interfaces, and the deployment details that make a project useful outside a notebook.
English · Tamil · Hindi (read & write)
SRM SDG Hackathon · SRMIST
BITS TechXcelerate Hackathon · BITS, Hyderabad
NLH AI × Legal Hackathon · MIT
Credential
Credential
C, C++, Python · FIIT
SRMIST · Credential
Python, C, C++, Java, Dart (Flutter), TypeScript, JavaScript
React, Electron, FastAPI, Flask, SQLAlchemy, Tailwind CSS, TensorFlow, PyTorch, Firebase, MySQL, SQLite
Machine Learning, Deep Learning, Reinforcement Learning, Computer Vision, NLP, OCR, local LLM inference (Ollama)
Linux System Administration, UNIX, Git, Docker, RBAC Systems, Performance Optimization, Unreal Engine 5, AirSim
OPEN TO THOUGHTFUL COLLABORATION
For roles, research, projects, or an interesting problem that needs a builder, reach out using any channel below.