Available for Software Engineering Roles • Hyderabad, IND
Engineering scalable systems & Generative AI architectures.
I am Velpula Kamalnath, a final-year Computer Science student (2026) at KMIT and former Product Development Intern at Darwinbox.
Experienced in full-stack engineering with JavaScript/TypeScript, Node.js, Python, Java, and C++,
building sub-second RAG retrieval pipelines with LangChain and Pinecone,
and designing robust microservices backed by optimized MySQL and MongoDB databases.
Developed a learning path generation tool in TypeScript (Node.js) utilizing AI coding copilots to accelerate development, structuring roadmap models that sequence topics logically and curate relevant educational resources.
Engineered AI-powered automation tools for the beacon.li platform using TypeScript, implementing UAT test actions to validate workflow behavior before release and documenting business requirements and technical specifications for engineering handoff.
Communicated progress and presented technical outcomes to stakeholders while collaborating cross-functionally with product and engineering teams on requirements, deployment planning, and feature reliability.
Architected a full-stack application following strict Model-View-Controller architecture, implementing client views and routing controllers in TypeScript (Node.js) while decoupling vector database query services.
Engineered the end-to-end AI/ML data pipeline in Python integrating LangChain and Pinecone for sub-200ms semantic similarity search over document chunks, backed by a MySQL (RDBMS) schema for persistent user sessions, query logs, and rate-limiting data.
Developed scalable TypeScript (Node.js) REST API endpoints with robust error handling, session validation, and unit/integration testing, reducing hallucinated responses by grounding replies in verified context.
Implemented custom TypeScript (Node.js) middleware for request sanitization, token validation, and response caching, optimizing query throughput and maintaining sub-second API response latencies under concurrent user loads.
Swaastha — Multi-Disease Health Screening Platform
Clinical AI Platform
Python (ML)TypeScriptReactNode.jsPyTorchTensorFlowMongoDBGemini API
Engineered a full-stack healthcare platform with a React (TypeScript) front-end, TypeScript (Node.js) backend service layer, and MongoDB database to persist multi-model prediction history and patient test records for longitudinal tracking.
Unified three AI/ML diagnostic models (CNN, ANN, Random Forest) implemented and trained in Python (PyTorch, TensorFlow) behind high-performance REST APIs, consolidating multi-modal inference data into a centralized diagnostic report.
Integrated Google Gemini API via structured prompt engineering in Python to generate automated, plain-language medical explanations, fusing ML model confidence metrics with LLM-driven diagnostic summaries.
Built secure TypeScript (Node.js) authentication with JWT tokens and optimized MongoDB aggregation pipelines to query user diagnostic trends, accelerating health dashboard load times and data retrieval efficiency.