I'm
Sudharsan Aravind
Backend Developer,Software Engineer,Machine Learning Engineer,
Full-Stack Developer
I'm a technology enthusiast with a background in designing and developing scalable applications from a young age. Along with an academic background in Data Science, I have a unique blend of strong foundation in Machine Learning and Software Engineering. I love to work in a collaborative team environment and also have good experience as both team player and team leader in multiple projects. My goal is to provide innovative solutions that are efficient and sustainable, contributing to the growth and success of an organization where I can apply my skills effectively while committing to both my personal and professional development.
Deep Forest Sciences | March, 2023 - March, 2024
One of the core developers of Prithvi, a next generation AI system that dramatically accelerates scientific discovery for human benefit.
PSG College of Technology | 2020 - 2025
CGPA: 8.27
SBOA Matric. and Hr. Sec. School | 2018 - 2020
Percentage: 88.5%
SBOA Matric. and Hr. Sec. School | 2017 - 2018
Percentage: 93.8%
Bank of Baroda & Microsoft Azure | July, 2024 - (ongoing)
Top 20 finalist among 21000+ participants in a Fin-Tech hackathon conducted by Bank of Baroda and Microsoft Azure. My team worked on a solution to provide personalized investment strategies and data-driven financial advice to the bank customers using Generative AI.
IIIT Sri City | December, 2023
Finalist in a health hackathon conducted by IIIT Sri City. Our team devised a solution to assist the patients to detect the symptoms of heart diseases using Machine Learning and Generative AI.
Ministry of Education, India | August, 2022
Winner and Team Lead of Smart India Hackathon 2022 held by AICTE. Our team built a semantic search engine for the Parliament of India and a built-in management portal to effectively maintain and supervise the Parliament records. I was the team lead and the solution architect of the project.
PSG College of Technology | January, 2022
One of the top 3 winners in a competitive coding contest conducted among participants from various computer science branches at university level.
Let's Code by KGISL Institute of Technology | October 2019
Creative Coder Award in Let's Code, a protoype building hackathon hosted by KGISL Institute of Technology.
Python | JavaScript | Flask | Node.js | React.js | Elasticsearch | PostgreSQL | sentence-transformers | AWS
EPACS - Effective Parliament Archive Connect System, a fully featured semantic search engine, tailor made for the Parliament of India as a part of Smart India Hackathon 2022 grand finale. The prototype aims to provide a comprehensive search engine for the Parliament records and a dashboard to manage the data pipeline. The project was initially intended to help eliminate the redundant questions asked in the Parliament. Later the project evolved to function as the primary search engine to retrieve the Parliament records.
Python | LangGraph | GPT-4o | FastAPI | Microsoft Azure | Cosmos DB | MongoDB | React.js | Deep Reinforcement Learning | DBSCAN
Smart Wealth is an AI-powered multi-agent financial advisor that delivers personalized, data-driven investment strategies using GPT-4o. It employs DBSCAN clustering to group investment options by market conditions, followed by a custom Deep Reinforcement Learning model to forecast and rank optimal strategies. The platform’s graph-based multi-agent system, built with LangGraph, includes specialized agents like the Market Analyst, Portfolio Planner, and Personal Finance Advisor, all coordinated by a Principal-Agent for comprehensive financial planning. Leveraging Microsoft Azure, Smart Wealth integrates Cosmos DB for metadata/vector storage, Blob Storage for documents, and Azure OpenAI for Large Language Models.
Python | Generative AI | Async-Programming | WebSocket | FastAPI | Qdrant | AWS | GitHub Actions(CD)
A Telegram bot that provides real-time press releases from the companies listed on Bombay Stock Exchange (BSE). The bot summarizes the press release documents and provides the user with the summarized content along with the type of the press release. The project is released as open-source under GPL-2.0 license and is currently at the alpha stage of development.
Python | Generative AI | Vision Transformers | FastAPI | PostgreSQL | Qdrant | Flutter
An AI powered platform that provides information about Ayurvedic medicinal plants for the Ayurvedic pharmaceutical industry. The platform uses a custom vision transformer model to identify the medicinal plants with high accuracy. At the heart of the platform is a chatbot - PlantGPT, that helps the users to interactively get information about the medicinal plants.
Python | YoloV9 | Amazon Textract | FastAPI | PostgreSQL | Async-Programming
Chequer is a modern and efficient solution for automating the cheque clearing process. It provides a streamlined workflow for processing cheques, reducing manual effort and improving accuracy. The system is built with a custom YOLOv9 model to detect the cheque details and Amazon Textract query service to extract the text from the cheque using natual language queries. At the heart of the system is an asynchronous queue system that provides a scalable and efficient way to process the cheques.
Python | Reinforcement Learning | Streamlit
A portfolio allocation system that uses Reinforcement Learning to optimize the portfolio allocation strategy. The system uses multiple RL agents to learn the optimal allocation strategy for the given market conditions.
© vsaravind.me. All Rights Reserved.
