Final-year IT Engineering student based in Pune, specializing in the intersection of Retrieval-Augmented Generation (RAG) and Knowledge Graphs.
My engineering focus is solving complex data propagation problems—ensuring that Large Language Models and backend systems retrieve from accurate, self-evolving truths rather than static, stale databases.
- Intelligence & Graphs: LangChain, LlamaIndex, PyTorch, Neo4j, Graph Neural Networks (GNNs)
- Backend Infrastructure: Python, Node.js, FastAPI, PostgreSQL, Redis, Docker, AWS
- Frontend & Experiential: React, Next.js, TypeScript, Three.js, MediaPipe
Self-healing knowledge graph framework. Engineered to monitor data sources and update graph nodes in real-time. Solves the core RAG bottleneck of stale data, ensuring high-fidelity, dynamically evolving retrieval for enterprise LLM agents.
Cascading state and dependency propagation system. Architected to map and manage the downstream impact of micro-changes within complex data ecosystems. Built to handle event-driven ripples across nodes, demonstrating deep systems-level state management and optimization.
Full-stack financial management platform. Built with a strict focus on robust data visualization, predictive financial analytics, and a clean, modular backend architecture handling complex numerical states.
Actively contributing to production-scale ecosystems, prioritizing substantive feature implementation and core infrastructure over superficial commits.
- Frappe (HRMS): Engineered core reporting logic, including the implementation of dynamic Return Amount columns in Employee Advance Summaries.
- Joplin: Resolved complex international keyboard layout handling and streamlined terminal application documentation.
- Ecosystem Engagement: Active technical participation and issue resolution across the Hugging Face Transformers and Twenty CRM ecosystems.
- The Tech Debt I Hate Most: "Tutorial clones" that don't solve real-world architectural problems.
- Current Learning: Advanced Graph Neural Networks (GNNs).
- Offline: Stealth mechanics (Hitman 3), T20 Cricket analytics, and Fitness.
"The best way to predict the future is to build it—then debug it."