Jul 2025 – Present · 11 mos · Bengaluru, Karnataka, India

• Kitchen Design AI Platform
• Architected and delivered a production-grade async API platform for complex AI workflows, implementing 202-accepted submission endpoints, long-polling result APIs, and clearly defined status contracts. Built end-to-end observability through structured logging with correlation-ID propagation across API, workflow orchestration, and CSP layers, enabling rapid debugging of distributed failures.
• Designed comprehensive validation and error handling with strict schema validation, explicit NKBA business-rule enforcement, and standardized error response models. Independently optimized a CSP-based layout engine, introducing grid-based spatial indexing, intelligent candidate ordering, collision-check caching, and fail-fast constraint evaluation—achieving zero quality loss while reducing solve time from ~30 minutes to ~3 minutes for high cabinet counts.
• Authored extensive technical documentation covering async workflows, message-queue orchestration, worker pools, PostgreSQL persistence, Redis Pub/Sub, WebSocket real-time updates, and NKBA rule interpretation strategies.
• Document Intelligence & OCR Platform
• Designed and implemented an end-to-end OCR and information-extraction pipeline, evaluating AWS Textract, Twix, MinorU, and Tesseract across accuracy, latency, and cost dimensions. Built a complete John Snow Labs NLP pipeline for structured data and PHI extraction from unstructured documents.
• Developed Temporal workflows orchestrating OCR processing, entity recognition, metadata creation, document chunking, vector store ingestion, and downstream querying—ensuring deterministic execution, automatic retries, failure isolation, and comprehensive observability across the pipeline. Delivered detailed OCR benchmarking analysis and technical documentation for JSL models covering ingestion, extraction, PHI handling, and consumption patterns.

Jul 2024 – Jun 2025 · 11 mos · Bengaluru, Karnataka, India

• EdTech Platform
• Architected and delivered core backend infrastructure for an EdTech platform serving 700+ students with 30K+ daily data transactions. Built standards-compliant REST APIs following oneEdTech specifications and implemented event-driven architecture achieving 99% processing reliability.Spearheaded performance optimization initiative, reducing critical database migration execution from 60 minutes to 5 minutes through query optimization, indexing strategy, and batch processing—enabling zero-downtime deployments. Designed and implemented serverless architecture using AWS Lambda, reducing infrastructure costs by 40% while maintaining sub-second API response times.
• Developed 10 REST APIs across Gradebook, Roster, and Resources domains, establishing API design patterns and validation framework adopted by 8+ team developers.
• Created real-time analytics dashboard with 9 AWS QuickSight widgets tracking student engagement metrics across 700+ interaction data points.
• Built robust event processing system handling 5 event types with automatic retry logic and dead-letter queue management.
• Performance Assessment System
• Automated quarterly performance review workflow, building PDF generation pipeline processing 100+ employee reports with dynamic templating and multi-language support. Implemented notification system with delivery confirmation, achieving zero missed feedback notifications across all review cycles.
• VibeCoders Marketplace
• Led technical evaluation for remote development infrastructure, conducting comprehensive analysis of GitHub Codespaces, Coder, and custom solutions across security, scalability, cost, and developer experience dimensions. Defined technical requirements for freelance marketplace including IP protection mechanisms, time tracking integration.

Jan 2024 – Jun 2024 · 5 mos · Bengaluru, Karnataka, India