Projects

• Created an AI agent capable of automatically selecting and utilizing appropriate tools to answer user queries.
• Developed a RAG using llama index library to enable the agent to reason over CSV and PDF data sources.
• Equipped the agent with functionalities for querying data, showcasing ability to interact with diverse tools.

• Built a clipboard manager using Python, rumps, and pyobjc, to access 100 clipboard entries from the menu bar.
• Reduced redundant copy-paste actions by 40%, significantly improving workflow efficiency.
• Packaged the app into a <10MB .dmg file using py2app for easy distribution and seamless user experience.

Developed a local context-aware chatbot with Python, LangChain, and Ollama's llama3.2 LLM to deliver privacy-focused and efficient conversational AI.

Created a Gmail filter using Google App Scripts to automatically label job rejection emails so that I do not get my hopes up with every new email.

Developed a Currency converter to ease the process of converting to INR every time I have to purchase something for home. Also added functionality for 8 more currencies using the API Ninjas API.

Created this portfolio website to showcase my profile in a better way to recruiters which included all my projects, accomplishments, blogs, etc in an easy to discover format.

• Optimized swarm robotics using behavior trees, reducing task time by 25% vs. centralized control.
• Implemented a foraging algorithm on a 49×49 grid with up to 14 robots, cutting messages exchanged by 50%.
• Designed and tested a collision-avoidance strategy for multi-robot coordination, in real-time movement.

• Created a Packet Sniffer tool working with protocol headers at datalink, network, and transport layer.
• Designed data structures for TCP, IP, ICMP and UDP headers and ran 3 experiments with automated 30s runtime.
• Parsed packet headers from raw sockets, mapping header fields to data structures based on protocol.

• Built a dumbbell network topology achieving accurate measurements across 5 experimental setups.
• Analyzed TCP Cubic and DCTCP performance by conducting 15 simulation runs and generating statistical results.
• Automated extraction and CSV generation for throughput and flow completion, ensuring standardized output.

Developed a Railway Ticketing Website using Ruby and Ruby on Rails and implemented features like search, book, and review tickets for passengers along with additional features for the admin.

The real world has a ton of data. Most of it is useless. Only a fraction of that huge chunk of data is useful and insightful. A Fishing Algorithm helps us "fish" out changing which x values gives the most change in the goal or y values. In short we developed a multiobjective semi-supervised explanation algorithm.

Created a Job search portal using Angular, HTML, CSS and JavaScript and implemented features like Search, Profile, Dashboard, Create and Apply for jobs, etc.