AI Engineer building scalable ML systems, real-time pipelines, and intelligent products.
Focused on LLMs, MLOps, and production-grade AI.
Still figuring out the meaning of life… but shipping models meanwhile 🚀

• Built a Deep Q-Network (DQN) agent to optimize traffic signal control using the SUMO simulator
• Modeled traffic control as a Markov Decision Process, learning signal policies from real-time state transitions
• Implemented experience replay + reward optimization to minimize vehicle wait times and congestion
• Integrated simulation control via traffic state representation and action-space phase switching
• Designed modular components for training loop, environment interaction, and policy evaluation

• Developed an agentic research system capable of multi-step reasoning and information synthesis
• Designed pipeline for query decomposition → retrieval → summarization → refinement
• Integrated LLMs with tool usage to simulate autonomous research workflows
• Supports iterative reasoning loops for deeper context building and answer quality improvement
• Structured for extensibility into fully autonomous AI agents / copilots

• Built a RAG pipeline for querying course/content data using semantic retrieval
• Implemented embedding-based search + context injection for accurate LLM responses
• Designed efficient document chunking and retrieval strategies to improve relevance
• Combined vector search + generation layer for grounded, hallucination-resistant outputs
• Structured for scalable knowledge base expansion and multi-domain adaptation

• Developed a general-purpose automation framework for orchestrating AI-driven workflows
• Designed modular architecture supporting task chaining, tool invocation, and execution pipelines
• Enables integration of APIs, scripts, and models into autonomous execution flows
• Focused on reducing manual intervention through programmable intelligence pipelines
• Acts as a foundation for building end-to-end agent systems and productivity tools