Sai Praveen
Tatiparthi

Software Engineer · San Jose, CA

• Joined a five-person team as an early engineer, owning the cloud deployment platform end to end for 2,600+ users: React frontend, TypeScript/Node.js/Express backend, PostgreSQL data layer, Docker, and Azure infrastructure as code (Bicep).
• Built a node-based visual workspace (React, ReactFlow, Zustand, Tailwind CSS) with 21 interactive modules, code editor, deployments, environments, logs, cost tracking, AI assistant, replacing dashboard navigation with a single canvas.
• Cut warm-path deployment time (cache-hit builds) from about 6 minutes to under 12 seconds, roughly 30x, across 100 to 150 deployment attempts a day, with a five-layer caching architecture and an autoscaled pool of pre-provisioned builders (13 ms median queue wait).
• Shipped the AI generation pipeline end to end on LangChain: durable, resumable generation jobs over SSE, RAG-grounded code generation over project state, multi-provider LLM fallback routing, and metered token billing with a free tier and per-tenant quotas.
• Built the engineering team's internal agentic harness: MCP servers and agent skills that give every developer's coding agent direct access to the platform's own context and conventions, standardizing how the team works with agents day to day.
• Built sandbox-backed live previews (e2b) so users see the real generated app running before any deploy.
• Engineered the reliability layer: a diagnose-and-repair agent auto-healing failed deployments within a bounded retry budget, tiered verification gating every deploy, and a 12-case generation-quality harness that verifies generated apps actually build in Docker and gates CI on a committed baseline; TDD across the stack, 1,900+ tests (Vitest, Playwright e2e) with coverage gates.
• Runs background processing on BullMQ/Redis job queues and Azure Service Bus messaging, WebSocket channels for live state, and an orphan-resource reaper that finalizes vanished deployments and cleans up cloud resources.
• Automates the Azure estate programmatically via ARM SDKs: Container Apps, Container Registry, DNS and CDN for custom domains with automated SSL, Blob Storage, and Cost Management.
• Instrumented with OpenTelemetry traces, Pino structured logging, Application Insights, Grafana dashboards, and alert rules as code.
• Made the platform commercially operable: Google/GitHub OAuth2 with JWT sessions, Stripe usage metering, and quota enforcement over per-tenant PostgreSQL isolation (Prisma).

Java Developer · India

• Architected full-stack features for finance web applications (Angular, Java Spring Boot) in an Agile/Scrum team using Jira and Confluence, improving transaction processing speed by 22%.
• Developed backend microservices for cross-service data synchronization, improving reliability by 28%.
• Contributed to AI-powered predictive analytics for financial risk assessment, 15% accuracy improvement.
• Integrated secure REST APIs with cross-functional teams, reducing client complaints by 12%.
• Automated regression testing with Selenium and Jenkins (40% less testing time); improved PostgreSQL queries and indexing (35% faster responses); implemented ELK-stack logging and monitoring.
• Led weekly code reviews and mentored interns, improving code quality scores by 27%.

Java Full Stack Developer · India

• Developed front-end modules for the company's e-commerce platform (AngularJS) with Java back-end services.
• Implemented Spring Boot REST APIs over MySQL for secure product and order management, with JUnit and Postman test coverage.