I am a Computer Science graduate student at Stony Brook University specializing in Data Science, with a strong foundation in machine learning and statistical modeling. I began by applying ML to signal processing and real-world engineering systems, then transitioned into NLP, working on transformer fine-tuning, discourse detection, and competitive retrieval tasks. Over time, I expanded into cross-domain ML applications including recommender systems, explainable AI pipelines, and distributed training. Currently, I focus on retrieval systems and Agentic AI, building context-aware, RAG-based architectures and participating in hackathons to develop scalable, production-ready AI solutions under real-world constraints.

Machine Learning research and applied AI engineering, with a strong focus on frontier applications of Agentic AI. I am particularly interested in designing autonomous, context-aware systems that improve existing workflows through intelligent orchestration, retrieval augmentation, and decision-making optimization. I aim to explore scalable ML systems, retrieval-enhanced generation, and automation frameworks that enhance real-world productivity and system efficiency.

RepSense, a hackathon project, is an agentic AI workout assistant that analyzes user workout data to provide adaptive, data-driven coaching and personalized performance insights. My university research focuses on improving retrieval systems using Geometric Modeling and Agentic Approaches, building upon existing dense retrieval and cross-encoder re-ranking systems and optimizing ranking performance through fine-tuning and systematic experimentation.

I’m a Computer Science graduate student from India exploring distributed systems and AI. I’m drawn to distributed systems not because I’ve mastered them, but because I haven’t — the failure modes, consensus mechanics, and recovery patterns are what make them exciting. I love building systems that coordinate, replicate, and scale. Recently, I’ve started experimenting with AI — especially LLM wrappers and autonomous task loops — with the goal of combining intelligent agents and backend infrastructure. Beyond coding, I’m a self-proclaimed cinephile and someone who believes in recalibrating when things get overwhelming (hence my Spotify playlist, “come back to yourself in whatever way you can”). I’m building, learning, and enjoying the process.

I’m deeply interested in distributed systems — consensus protocols, replication, fault tolerance, and scalable backend architecture — and I’m eager to keep learning everything I can in this space. I’m also exploring AI, especially how LLMs and autonomous agents can integrate with infrastructure and observability systems. I’d love to connect with engineers working at the intersection of distributed systems, cloud platforms, and intelligent tooling.

Right now, I’m tinkering with two things: a Raft-based distributed key-value system where I experiment with failure injection, replication tuning, and observability under load, and a lightweight Claude wrapper that runs autonomous task loops — repeatedly executing structured prompts, evaluating outputs, and iterating until a convergence condition is met. I’m exploring how distributed infrastructure and AI agents can work together to build systems that not only scale and recover, but also reason.

I started coding in 8th grade, driven by curiosity and a habit of learning by building. By the time I entered university, I was already chasing real engineering problems, which helped me crack competitive internships early on. In one case, I was selected as one of just 2 candidates out of nearly 400 applicants, an experience that pushed me to raise my standards even further. Over time, I built a reputation in my university as someone who could take on hard systems problems and actually ship. From distributed systems and backend engineering to now exploring agentic AI and intelligent applications, I’ve consistently gravitated toward ambitious, technically demanding work that sharpens both depth and taste.

I’m interested in building AI systems grounded in distributed systems principles, focusing on reliable agent orchestration, state management, fault tolerance, and scalable inference. I want to apply concepts like consensus and replication to design robust, production grade LLM applications and multi agent systems.

I’m focused on AI systems, especially agentic architectures that can reason, use tools, and handle multi step workflows. I’m exploring orchestration frameworks like LangChain, memory design, retrieval, and evaluation to build intelligent applications that go beyond simple prompt responses. I’m interested in production grade LLM systems that are reliable, scalable, and thoughtfully designed end to end.