Hey, I'm
Arya Pathak
Computer Engineering @ VIT Pune · SDE Intern at PhonePe · Go developer · Open source contributor ·
Where I've worked
Designed a bulk gift card creation API eliminating manual ops handoff. Built maker-checker dashboards with Fabricator for ops team autonomy. Redesigned EGV redemption workflow using shard-level batch fetches to support 300+ cards; migrated OkHttp to PhonePe's internal HTTP handler. Enhanced platform reliability by modernizing dependencies, enabling YAML anchor reuse (Rosey), refactoring JUnit with Guice injection, and rolling out Feature Factory CI across PPI projects (Wallet, Gift Card Service, Distributor Service).
- Designed and built a bulk gift card creation API for Qwikcilver in a Java Dropwizard microservice, exposing 5 endpoints for synchronous B2C single creation, async B2B bulk creation, cancellation, status, and details, eliminating the manual ops handoff previously required to provision gift cards.
- Built maker-checker dashboards for the bulk GC creation flow using Fabricator (internal drag-and-drop dashboard tool), enabling ops teams to initiate, review, and approve bulk gift card requests without engineering intervention.
- Redesigned the EGV redemption workflow to replace per-card DB queries with shard-level batch fetches after balance reservation, supporting users with 300+ gift cards without degrading performance; migrated OkHttp calls to PhonePe's internal HTTP handler in the Gift Card Service.
- Modernized dependencies and enabled YAML anchor reuse in service configs (Rosey); refactored JUnit test infrastructure with Guice-based injection and shared base classes to cut configuration duplication; rolled out Feature Factory CI support across all PPI projects (Wallet, Gift Card Service, Distributor Service) to automate feature environment creation in GitHub pipelines.
Built a Golang AI agent bot on Microsoft Teams using Gemini API to automate GCP IAM tasks via natural language. Deployed a Cloud Run backend for bulk onboarding and audit logging — cutting setup time by 45%.
- Built and deployed a Golang-based AI agent bot integrated with Microsoft Teams to automate GCP onboarding and account management via Gemini API function calling, enabling non-technical staff to perform IAM, role, and service account operations through natural language.
- Designed a secure, cloud-native backend on Google Cloud Run supporting bulk onboarding, role assignments, and audit logging — reducing multi-user setup and query resolution time by 45%.
Built a sandboxed in-browser Python execution engine using WebAssembly and Pyodide for 500+ users. Deployed a Node.js/Express RCE backend on AWS EC2, and published a C# NuGet package that cut execution latency by 30%.
- Built an in-browser Python execution environment leveraging WebAssembly and Pyodide, enabling real-time, sandboxed code execution for 500+ game users across web, PC, and Mac platforms.
- Engineered and deployed a secure Python Remote Code Execution engine using Node.js and Express.js on AWS EC2 with Nginx reverse proxy, processing 1K+ executions/month for 500+ users via WebAssembly and Pyodide.
- Published a C# NuGet package for safe local Python code execution, reducing latency by 30%.
Things I've built
Lightweight monitoring service in Go using the Chi router and PostgreSQL, supporting 100+ resources with persistent state. Real-time notifications via Pusher WebSockets and 500+ scheduled health checks daily via Cron, delivering alerts in under 2 seconds.
Serverless, decentralized file sharing platform in Go enabling direct P2P transfers over TCP without a central coordinator. AES encryption for files at rest and in transit with concurrent, encrypted streaming and Gob-based message encoding.
Modular desktop packet sniffer with a responsive Fyne GUI in Go, using C and cgo for efficient cross-platform packet capture and live protocol inspection. AI-powered natural language filtering via Google Gemini with gopacket for dynamic plugin-based enrichment.
Flask backend connecting WhatsApp with Google Workspace to automate scheduling, Drive uploads, and Meet workflows while processing 100+ messages/day. Uses OpenAI-powered NLP, regex, and observability tooling for high intent-detection accuracy and real-time monitoring.
Cloud-native online assessment platform with ASP.NET Core microservices, RabbitMQ, Angular, and a secure Golang code-execution engine. Built for scalable delivery, automated grading, and containerized deployment on Kubernetes and Azure.
Computer vision project implementing a Pix2Pix GAN to translate satellite imagery into map-style outputs. Built with PyTorch using U-Net and PatchGAN components.
Handwritten compiler inspired by Roslyn with recursive descent parsing, syntax trees, diagnostics, semantic binding, REPL execution, and xUnit-backed test suites.