Software engineer with 6+ years architecting cloud-native, event-driven microservices on AWS that run mission-critical operations at enterprise scale. I own systems end-to-end, from ambiguous business goals to production, that prevent tens of millions in operational losses, serve 20M+ transactions daily, and cut engineering toil through automation and Gen-AI tooling. Currently a Software Engineer at Amazon, based in Seattle, WA.
- 🔭 Deep expertise in distributed systems, serverless & streaming architectures, and Infrastructure as Code
- 🤖 Building with Gen-AI tooling : MCP servers, agentic hooks, spec-driven development, LLM ops
- ⚡ Bias for high-leverage initiatives, operational rigor, and raising the technical bar across the team
- 📫 Reach me: agsai0128@gmail.com
I specialize in tier-1, always-on systems, owning them end-to-end from architecture to production to on-call. I care about distributed systems that hold up under scale and about the kind of automation that compounds, making everyone on the team faster, not just me.
University at Buffalo — MS, Data Science | New York, USA
Nitte University — BS, Electronics & Communication | Bangalore, India
SWE @ Amazon | May 2022 – Present | Seattle, USA
- Owned technical strategy for a C-Suite goal to cut outbound concessions — designed and shipped an event-driven defect-detection service that flags incorrect items before they leave the facility, eliminating 75% of manual inspection and preventing $45M in annual concessions
- Designed and deployed a serverless control service for an Automated Storage and Retrieval System (Lambda, API Gateway, S3) that interfaces with warehouse hardware, cutting per-item processing time 30% and lifting order throughput 50%
- Engineered and optimized 4 tier-1 microservices in Java & TypeScript, including a hazmat/battery compliance service preventing $100K-per-package fines and a label-rendering service processing 20M+ shipping labels daily at 25% lower latency
- Led a team-wide initiative embedding Gen-AI tooling across 90% of the dev lifecycle — stood up internal MCP servers and agentic hooks (code gen, test scaffolding, PR review), cutting manual coding time 40% and accelerating delivery ~20 days per release
- Coordinated 24/7 on-call for 50+ high-severity, region-wide incidents/year, reducing MTTR 35%
Software Engineer @ Amigo Sport | Sept 2019 – Nov 2020
- Built a real-time player detection & tracking system (Python, YOLOv2, OpenCV) that classified video frames by player coordinates to generate formation labels, eliminating 40 hrs/week of manual labeling
- Implemented a spatial detection model achieving a 3.5cm error rate and sub-2s latency, meeting FIFA accuracy standards for real-time officiating support
- Developed an action-recognition & image-classification pipeline (Python, TensorFlow, OpenCV) driving $10K/month in new revenue
Systems Engineer @ Infosys | Nov 2017 – Jun 2019
- Architected large-scale data-migration pipelines on AWS (EC2, Route 53, S3, RDS, ElastiCache, IAM, Glue), migrating 450M+ records into Oracle Flexcube with zero data loss and cutting cycle time 30%
- Modernized Citibank's legacy banking systems onto Oracle Flexcube with RESTful APIs (ISO20022) + SAML 2.0, improving availability to 99.95% and supporting auto-scaling for 5M+ daily transactions
Event-Driven Defect Detection : Real-time service that flags mis-shipped items before they leave the facility. Event-driven architecture on AWS; prevents $45M/year in concessions.
Serverless ASRS Control Service : Elastic Lambda/API Gateway/S3 service interfacing directly with warehouse hardware; cut per-item processing time 30% and lifted throughput 50%.
Internal Gen-AI Developer Platform : MCP servers exposing codebase, docs, and CI/CD context to AI agents, plus spec-driven development and agentic hooks embedded across 90% of the dev lifecycle.
Languages & OS
Cloud & Distributed Systems
Streaming & Data Pipelines
DevOps & Testing
Gen AI
