Hi, I'm Prathamesh Chaudhari 👋

Software Engineer · Android × Full-Stack × Cloud/DevOps
CS grad @ VIT · SWE @ Avegen Health · First-author IEEE author

• 🎓 B.Tech in Computer Science & Engineering @ Vellore Institute of Technology (2022–2026)
• 💼 Software Engineer @ Avegen Health, shipping production Android + full-stack systems for clinical-research platforms
• 📄 First-author IEEE publication on deep-learning-based land-cover change detection
• 🌱 Open-source contributor to autonomous-AI-agent infrastructure
• ⚡ I like owning things end-to-end — from a Kotlin UI down to the CI/CD pipeline and cloud infra that ships it

A distilled view of the engineering I do day to day — built up across production work at Avegen Health.

• 📱 Production Android (Kotlin/Compose) — on-device analytics telemetry, Microsoft Entra (MSAL) SSO, Firebase Crashlytics, encrypted Room/SQLCipher storage, and multi-tenant build-flavor matrices
• 🌐 Full-Stack TypeScript — Next.js (App Router) + Prisma/PostgreSQL backends, REST API design, full-stack auth features, and static-export marketing sites with serverless functions
• ☁️ Cloud & DevOps on Azure — Container Apps, Bicep IaC, Azure DevOps pipelines, idempotent blue-green deployments, scheduled cost-control automation, and Static Web Apps
• 🔐 Security-Minded Engineering — SSL certificate pinning (VAPT-hardened), hashed/expiring tokens, rate limiting, PII/PHI isolation, and secrets managed via Azure Key Vault
• 🧪 Testing & Reliability — cross-service E2E with Playwright + Maestro, JUnit/Robolectric, and hardening flaky CI against emulator cold-starts
• 🔧 Build-Toolchain Modernization — coordinated JDK / AGP / Kotlin / Gradle upgrades and the dependency migrations they require

Languages

Frameworks & Libraries

Cloud & DevOps

Databases & Testing

ML / AI

Open-source autonomous AI agent framework

Contributing reliability fixes, new capabilities, and hardening across the agent runtime, gateway, desktop, messaging adapters, and checkpoint/tooling layers.

Open-source AI software-development agent

Contributing to the app-server and deployment layer — reverse-proxy traffic handling, memory-integrity processing, and a security fix for command injection in git config.

Land Cover Change Detection using DeepLabv3-ResNet50 — first author 2026 4th Intl. Conference on Intelligent Data Communication Technologies and IoT (IDCIoT), IEEE, 2026. 📎 IEEE Xplore · DOI: 10.1109/IDCIOT67589.2026.11455786

• Claude Certified Architect — Anthropic 🔗 Credly
• GDPR Compliance — Udemy 🔗 Certificate
• HIPAA Training — Udemy 🔗 Certificate

A launch-focused product workflow that lets product teams pressure-test messaging, pricing, and market risk before shipping. It turns a structured product brief into a knowledge graph, expands the audience into archetype-driven consumer segments, and runs three scenarios in parallel — baseline, competitive pressure, and crisis response — then surfaces sentiment movement, objections, faction shifts, and archetype heatmaps in a single dashboard.

Highlights

• 🧠 Brief → knowledge graph → archetype-based population synthesis (LLM + Zep)
• ⚡ Parallel campaign execution with async, poll-tracked preparation tasks
• 📊 Per-simulation sentiment caching with log-driven invalidation
• 🔒 Locked-down production defaults — required SECRET_KEY, debug off, origin-based CORS
• 🐳 Single-image Docker deploy: backend builds + serves the compiled frontend from one process

Tech: Python · Flask · Vue · LLM (OpenAI-compatible) · Zep knowledge graph · Docker