Bengaluru, Karnataka, India
Full-stack engineer (5+ yrs), backend-focused — build config-driven platforms, APIs, and auth infra in Python / protobuf / Node with React front ends. Own features end-to-end, from low-level design to production, with a focus on clean, reusable code and reliable rollouts.
- Built the Enterprise Search configuration platform (React + Python) spanning 10+ SaaS connectors, fully configurable UI and backend for all connectors — adopted by 300+ enterprise customers. - Created an admin tool that loads proto-backed configs as editable JSON and saves them back across 200+ config types — letting partners configure UIs faster, bulk-upload data, and onboard new customers quickly. - Designed web session + logout infrastructure end-to-end (Python/protobuf auth service, Node BFF, client SDK), securing sign-out and SSO revoke across 3 apps in all environments.
- Developed the BFF for Moveworks' protobuf-based configs generic fetch/save handlers and a schema-driven framework powering 50+ self-serve admin pages with form generation, validation, and patch updates. - Led backward-compatible, zero-downtime schema migrations (field-to-field dual-writes, backfill scripts) and drove refactors improving readability, reusability and reliability.
- Worked end-to-end across Eightfold's Talent Acquisition dashboard and built the candidate profile page, cutting its initial load time ~50% by parallelizing previously-serial backend calls and lazy-loading heavy sections — React, Python/Flask, Solr.
- Built the Inbox app — recruiter↔candidate email/SMS across Inbox/Sent/Scheduled tabs, with conversation-based email threading and polling-with-backoff to keep messages fresh — Python/Flask backend, JavaScript UI. - Implemented scheduled email delivery with a store-and-dispatch design — a periodic worker pulls due emails from a DB (by send_at) and dispatches them through AWS SQS.
- Improved search-suggestion relevance for the Outlook web app (ML team): analyzed user interaction data (past searches, clicks) and trained an ML model to predict the most-likely-clicked suggestion, sharpening how query-formulation suggestions were ranked | Python / Jupyter.