CV

• Led redesign and deprecation of a 17-year legacy Tier-1 distributed configuration platform handling 10M+ daily requests across 200+ global fulfillment sites, consolidating fragmented schemas into unified DynamoDB-backed APIs with zero production downtime.
• Designed and implemented large-scale migration infrastructure including schema redesign, migration tooling, double-read/double-write validation frameworks, staged traffic shifting, rollback mechanisms, and production monitoring.
• Developed distributed backend services and APIs in Java/Spring using DynamoDB, ECS, Lambda, SQS, SNS, Kinesis, and CI/CD tooling while driving tradeoff decisions between scalability, fault-tolerance, and blast radius minimization.
• Owned operational reliability for 3+ Tier-1 production systems including deployment pipelines, dashboards, alarms, incident response, load testing, and production migration coordination.
• Resolved critical European-wide production incidents by identifying transient distributed bugs and introducing safeguards across 3+ Tier-1 services.
• Designed and shipped 3 key features across 8+ systems, earning/saving estimated $450M per year — focused on reducing manual intervention and scaling across more sites.

• Co-author and co-lead implementer of a state-of-the-art robotic tabletop rearrangement system published at ICRA 2025, focusing on long-horizon planning and robust execution under physical-world uncertainty.
• Built planning systems integrating vision-language models and large language models for mobile manipulation and multi-step reasoning tasks.
• Prepared study resources and provided office hours and seminars for a Search and Optimization course.

• Designed and implemented experimentation infrastructure integrating an in-house A/B testing backend with a high-throughput widget delivery platform serving 5k+ RPS.
• Built backend APIs, integration services, rollout tooling, and React-based frontend systems enabling rapid experiment configuration — contributing to £150M+ YoY business impact.
• Designed backend integrations between experimentation systems, widgeting services, and analytics platforms using TypeScript, Python, SQL, React.js, and MVC architectures.
• Built SQL and Python pipelines for experiment evaluation, statistical analysis, and operational analytics across millions of daily events.
• Managed production infrastructure including Kubernetes clusters, deployment pipelines, VM provisioning, patching, and operational maintenance.

OmniManip: Towards General Robotic Manipulation via Object-Centric Interaction Primitives as Spatial Constraints

ICRA 2025

Co-author; contributed to long-horizon robotic planning, manipulation systems, and LLM-assisted embodied AI reasoning.

In Search for Optimal Schedule During Long-Term Treatment of Port-Wine Birthmarks

Scientific Reports

Co-author; applied machine learning and statistical analysis methods to medical outcome prediction from imaging datasets.