Specializing in Agentic Workflows, Graph Algorithms, and Large-Scale RAG Architectures.
I bridge the gap between theoretical machine learning and scalable production systems. My engineering philosophy centers on optimizing compute economics and solving high-concurrency constraints.
Currently, I am an AI Engineer at Wappnet Systems, where I was awarded Employee of the Year (2024) for architecting enterprise AI solutions. My work focuses on:
- Graph Theory: Optimizing pathfinding on large-scale graphs ($>$50k nodes) using Contraction Hierarchies.
- Agentic AI: Building autonomous multi-agent systems using the Model Context Protocol (MCP) and OpenAI Agents SDK.
- Deep Learning: Researching hybrid architectures (LSTM+CNN) for time-series forecasting on satellite imagery.
| Domain | Stack |
|---|---|
| Languages | |
| AI & ML | |
| Backend & Systems | |
| Data Engineering |
- AI Career Counselor: [+1 (775) 640-9474](tel:+1 (775) 640-9474)
- A real-time voice agent capable of autonomous research. Built with Twilio & OpenAI to demonstrate accessible, privacy-first personal assistance over standard telephony—no internet or apps required.
- rustlette: A high-performance Rust implementation of the Starlette ASGI framework. Focused on zero-cost abstractions for async web services.
- fastapi-smith: An opinionated project generator (CLI) for production-ready FastAPI architectures. Available on PyPI. View docs here
- calorie-tracker-ai: "Snap & Track" nutritional analysis engine. Combines Computer Vision for object detection with Multimodal LLMs to estimate caloric content from single images in real-time.
- teleai.tech: Enterprise AI Telephony platform.
- voice.teleai.tech: Real-time Voice-to-Text conversational interface. Engineered on the HCI principle that humans speak faster than they type, but read faster than they listen. Features instant low-latency generation for rapid information retrieval.
An intelligent AI-powered chat assistant with seamless integrations
- AI-powered chat with context awareness
- Gmail integration for email management
- Slack integration for team collaboration
- Jira & Notion task management
- Calendar integration (Outlook & Google)
- Intelligent search across all platforms
- Secure OAuth authentication
- Live Application: chat.teleai.tech
- Visit chat.teleai.tech
- Sign up using your email or OAuth provider
- Connect your integrations (Gmail, Slack, Jira, Notion, etc.)
- Start chatting with your AI assistant!
- Frontend: React, Vite, TailwindCSS
- Backend: FastAPI, Python, PostgreSQL
- AI: OpenAI, LangChain
- Infrastructure: AWS EC2, Nginx, PM2
- CI/CD: GitHub Actions
- The Challenge: Calculating optimal sea routes on a 50,000+ node graph with sub-second latency constraints.
- The Innovation: Implemented Contraction Hierarchies to pre-process graph overlays, moving beyond standard Dijkstra/A* limitations. Re-architected the Python request handler to solve GIL (Global Interpreter Lock) bottlenecks.
- Result: Reduced query latency from 2s to <100ms (p99).
- Stack: Python, Graph Theory, Threading.
- Organization: Indian Space Research Organisation (ISRO).
- The Challenge: Forecasting solar events using 200GB+ of multi-sensor satellite raster data.
- The Innovation: Developed a hybrid LSTM + CNN architecture to capture both spatial features (CNN) and temporal dependencies (LSTM).
- Result: Outperformed baseline ARIMA/SARIMA models by 40% during high-fluctuation weather events.
- Resources: 📄 Read Research Report
- Stack: PyTorch, OpenCV, SciPy, CUDA.
- The Challenge: Preventing infinite reasoning loops in autonomous AI agents while managing costs.
- The Innovation: Built a multi-agent state machine using Redis for active memory (<5 min window) and PostgreSQL for long-term recall. Developed custom MCP Servers for Slack/Jira integration.
- Result: Reduced token consumption by 63% via prompt optimization and separated "Reasoning" vs "IO" agents.
- Stack: FastAPI, OpenAI Agents SDK, Redis.
- Employee of the Year (2024) | Wappnet Systems
- Awarded for rapid delivery of high-impact AI architectures across Healthcare & Logistics sectors.
- Smart Kart Innovation Project | _VOIS Innovation Marathon | View Certificate
- Developed an IoT-enabled Smart Kart prototype during the Vodafone Idea Foundation internship.
- Supervised Machine Learning: Regression and Classification - DeepLearning.AI & Stanford University | View Certificate
- Instructor: Andrew Ng. Focus on gradient descent and logistic regression architectures.
- Python for Data Science and Machine Learning Bootcamp - Udemy | View Certificate
- Comprehensive curriculum on Scikit-Learn, Pandas, and Neural Networks.
- GDSC Core Team Member (2021-2022) - Google Developer Student Clubs | View Certificate
- Recognized for technical leadership and community building at New L.J. Institute of Engineering and Technology.
Open to research collaborations and engineering opportunities.