I'm Urvashi Kohale, a Software Engineer and M.S. Software Engineering student at San José State University, graduating in December 2026. I enjoy building software across the stack from responsive user interfaces and scalable backend APIs to distributed systems, cloud infrastructure, data pipelines, and AI-powered applications.

I'm currently seeking full-time new-grad Software Development Engineer opportunities and am open to full-stack, backend, platform, infrastructure, cloud, data, and AI/ML engineering roles.

• 💻 Build end-to-end applications using React, Next.js, FastAPI, Spring Boot, Node.js, and modern databases
• 🔗 Design reliable APIs, microservices, distributed systems, and event-driven architectures
• ☁️ Deploy and operate software using AWS, GCP, Azure, Docker, Kubernetes, Terraform, and CI/CD
• 📊 Develop streaming pipelines, database-backed services, monitoring, and performance optimizations
• 🤖 Create RAG, multimodal AI, machine-learning, and intelligent full-stack applications
• 🤝 Collaborate across teams, translate ambiguous requirements, and deliver production-ready solutions

A GFS-inspired distributed storage platform designed for concurrency, consistency, and fault tolerance.

• Implemented a metadata server, chunk servers, replication factor of three, heartbeats, and failure recovery.
• Added multithreaded reads and writes, per-chunk locking, strict lock ordering, and lease-based consistency.
• Designed the system to support more than 100 concurrent client requests.

A high-throughput streaming pipeline for ingesting, processing, and storing real-time events.

• Built a pipeline processing approximately 30,000 events per minute.
• Reduced end-to-end latency from approximately four seconds to under one second using batching and bulk inserts.
• Added consumer-lag monitoring, offset management, backpressure handling, and scalable consumers.

A multimodal, agentic AI platform that converts educational documents into interactive video content.

• Developed a document-to-video workflow combining retrieval, content generation, and media processing.
• Used FastAPI, Celery, and Redis to execute long-running tasks asynchronously and in parallel.
• Designed retrieval workflows using LangChain and ChromaDB for contextual grounding.

A full-stack vision-language nutrition assistant that analyzes meals and provides grounded nutrition insights.

• Built an interactive React and FastAPI application for image-based meal analysis.
• Evaluated GPT-4o, Qwen-VL, and LLaVA for multimodal food understanding.
• Grounded nutrition estimates using verified USDA FoodData Central information.

A full-stack event discovery and ticketing application with secure payments and scalable services.

• Developed user-facing event discovery, booking, authentication, and payment workflows.
• Built REST services with Spring Boot and persisted application data in MongoDB.
• Automated builds and deployments through a Jenkins-based CI/CD pipeline.