Projects

Description

A Kubernetes-native Go runtime for crash-resumable, multi-agent LLM workflows with durable execution, distributed scheduling, and event-sourced observability.

Achievements

• Added per-step checkpoints so long-running workflows can resume after crashes without duplicating tool calls or model spend.
• Scheduled DAG-based workflows across a scalable Kafka-backed worker pool with per-task retries.
• Captured every model and tool call through event-sourced tracing, audit logs, and real-time workflow monitoring.

Description

A Python MCP memory harness that carries decisions, project facts, and preferences across AI coding sessions while detecting stale or conflicting context.

Achievements

• Reduced repeated context tokens by approximately 40% through cross-session state handoffs.
• Consolidated raw sessions into durable short- and long-term memory with stale-memory cleanup and conflict detection.
• Validated retrieval reliability with Recall@5, MRR, and traces for ranking regressions and injected context.

Metrics

• ~40% fewer repeated context tokens
• Recall@5 and MRR evaluation

Description

An MCP-powered Figma agent built for the Microsoft AI Hackathon, using a LangGraph supervisor and ReAct tool loops to turn natural-language prompts into FigJam components.

Achievements

• Converted natural-language prompts into FigJam components at approximately 85% first-try accuracy.
• Reduced API costs by approximately 28% and lowered latency with PostgreSQL-backed memory, caching, and model routing.

Metrics

• ~85% first-try accuracy
• ~28% lower API costs

Description

Optimized edge computing AI recognition using KubeEdge, streamlining local computations and decision-making. Devised a Convolutional Neural Networks model in TensorFlow for vegetable recognition, achieving 98.6% accuracy. Built a centralized management system using Django with MQTT, enhancing live oversight in multiple edge modules. Fostered system resilience and prompt updates by integrating a Kubernetes orchestrated CI/CD pipeline.

Description

Built an interactive web-based tool using the Flask web framework, along with Dash and Plotly for UI design, to analyze the temporal and spatial relationships between taxi trips and city popularity in Manhattan. Processed and cleaned 30.8 million taxi trip records using PySpark and improved regional analysis accuracy by integrating data from the Google Popular Times API. Managed databases interactions using SQL Alchemy, implemented a multi-regression model for estimating missing taxi trip coordinates, and developed a user-friendly interface with customizable filters for data analysis.

Description

Architected and developed a full-stack library management web application using React and Spring Boot with a RESTful API for seamless frontend-backend communication. Integrated Okta's Authentication API and Stripe's Payment API to provide secure user authentication and reliable processing of credit card payments for book late fees. Utilized Maven for streamlined dependency management and Spring Data JPA to interact with a MySQL database. Created a Books module with CRUD functions and enable features such as rating, commenting, loan/return books. Developed an AI chatbot assistance agent by utilizing OpenAI's GPT-3 API.

Description

Developed a comprehensive invoice management system with MongoDB, Express, React, and Node.js, concentrating on dynamic invoice generation, user administration, and real-time revenue monitoring. Utilized Docker Compose for streamlined, microservices-based deployment, boosting system scalability. Fortified user authentication via a refresh token approach, Google OAuth, and stringent email validation mechanisms. Enhanced system performance under high traffic by integrating NGINX for load balancing and reverse proxying. Applied Winston and Morgan for efficient log monitoring and anomaly detection.

Description

Build a cross-platform mobile messaging app with React Native, Expo, and Redux, incorporating key features such as instant messaging, user search, group chats, and custom profiles. Designed a scalable, distributed system using Firebase's NoSQL for real-time data synchronization to maintain secure authentication and end-to-end encryption for data integrity. Create a context-sensitive AI voice interface using OpenAI's GPT-3.5 Turbo, Whisper API, and Azure TTS API.