Hands-on engineer and technical leader with 12+ years of experience leading critical distributed systems, and turning domain problems into company-wide platforms. Over nine years at Amazon Prime Video I built and led several core backend systems, including the edge routing platform for global traffic management and failover (millions of requests per second across eight AWS regions at 99.99% availability) and the customer location platform for 200M+ customers and 100+ teams. I pair this depth with company-wide influence: I aligned 10+ organizations & 50+ teams to one routing standard, wrote a chaos-testing framework adopted by 200+ teams and Amazon's official guide for writing technical design documents.
- Defined and drove the 3-year architecture roadmap for the global edge-routing platform, which fronts 900+ APIs from 50+ service teams across 10+ organizations.
- Cut route-change lead time from one month to one day by redesigning the route-change workflow to remove a 20-stakeholder manual approval path, letting Prime Video grow live sports from tens to hundreds of events a year with fewer customer-impacting incidents.
- Technical lead for the multi-tenant platform that serves millions of requests per second at <10 ms p99 latency and >99.99% availability across eight AWS regions, with an 8-engineer team.
- Conceived and bootstrapped an agentic code-migration platform that reduced migration time from weeks to approximately one day and was adopted by 50+ Prime Video teams.
- Built the agent orchestration architecture, including specialized-agent workflows, evaluation frameworks, reusable prompt abstractions, custom tools, and context-management.
- Made chaos testing cheap enough to run continuously in any AWS-service pipeline: created an open-source fault-injection framework adopted by 200+ teams and integrated it into Amazon's company-wide resilience scorecard, so services prove they survive CPU, memory, disk and dependency failures before every release, and prevent outage-causing bugs such as a misconfigured dependency timeout from reaching production.
- Grew it from a one-team library into a company standard and owned it for six years, leading contributors from across Amazon; presented it at AWS re:Invent and received a VP-level above-and-beyond award.
- Authored Amazon's official guide for writing technical design documents, adopted by the central engineering training org and issued to new software engineers at onboarding.
- Published three technical deep dives through the official AWS Open Source and Networking blogs (1, 2, 3), and presented at multiple internal engineering conferences.
- Coached 15+ Amazon engineers through promotion to mid and senior levels; active interviewer on London hiring loops and mentor in student-outreach programs.
- Built and scaled Prime Video's customer location platform to millions of RPS, delivering globally replicated customer location data with regulatory compliance. A hard dependency for 100+ teams.
- Built Prime Video's Client Notifications platform - the async device notification layer that let partner teams ship live sports UI updates and the Watch Party feature to millions of customers. Sub-minute fan-out to 10M simultaneously connected devices across living room and mobile apps.
- Re-designed a cross-region DynamoDB global replication system, replacing the previous design: availability went from 99% to 99.999%, replication latency 2s to 200ms and outperforming AWS's own Global tables.
- Re-architected the cache eviction layer for Prime Video's location service - improving cache hit rate from 80% to 99% and saving $1M/year in database costs.
- Second engineering hire on a four-person founding team; built core systems for a photo-based fashion-shopping platform, including an image-vectorization pipeline for affiliate-product detection.
- Re-architected the database layer, reducing latency by 90%.
- First R&D engineer; built research systems for timestamp alignment, lyrics-based recommendation, and automated improvement of user-annotated lyrics.
- Published viral data studies featured by news outlets, such as The Guardian and represented the company at 20+ hackathons in the US and EU.
Yamaha Music R&D, 2013 (Japan) · Google Summer of Code 2012, Mixxx.org (Spain) · Google Summer of Code 2010, Beagleboard.org (India)