Mark – Python, Flask, FastAPI

Designed and evolved core services powering a high-scale payment infrastructure, including the API gateway, Webhooks delivery, Public API, Massive operations Service, Search service and many more. Part of Platform / DevEx team with a main goal of improving development velocity while having ownership on the most critical services that served as the primary entry point for all customer traffic and internal integrations, requiring high availability, low latency and strict reliability guarantees.

• Owned and evolved the API Gateway service — the single entry point for all public traffic — ensuring high availability, low latency and 99.99% SLO.
• Built and maintained the Webhook service, reliably delivering real-time event notifications to clients at scale with guaranteed delivery semantics with a scale of 300+ million events.
• Designed and shipped the Mass Ops service, enabling engineering and merchant operations teams to implement workflows execution using Temporal — reducing per-team integration effort.
• Improved organization-wide CI/CD performance by 40% through pipeline optimization, build parallelization and tooling improvements.
• Led engineering adoption of AI-assisted development workflows, integrating MCP-powered tooling, Claude and CodeRabbit into the development lifecycle.
• Collaborated with multiple engineering teams to define platform standards and improve operational excellence.

Greenfield project. Architected and built a cloud-native video streaming platform from scratch for smart camera devices. The system enabled live video streaming, motion-triggered recordings, event processing and cloud-to-device communication for large-scale IoT deployments.

• Architected and built a WebRTC video streaming platform from scratch as sole backend engineer, delivering a production-ready system end-to-end.
• Designed the video streaming supporting real-time viewing and recording workflows.
• Designed and owned the events service, capturing motion and on-demand recording events with low-latency access patterns.
• Took full ownership of cloud infrastructure (AWS, Terraform) — provisioning, scaling and maintaining all platform resources from day one.

Built and scaled the core machine learning data platform used for training computer vision models powering Ring cameras and smart home devices. The platform supported video ingestion, annotation workflows, dataset generation, secure access management and AI-assisted labeling.

• Designed, scaled and took under ownership 10+ platform's core video and annotation services supporting millions of training assets.
• Built and scaled the core video & annotation service for ML data platform, achieving an 4× performance improvement by profiling bottlenecks and redesigning critical data access patterns to handle significant RPS.
• Redesigned the dataset management service — reduced dataset generation time from 12 hours to 5–10 minutes by rearchitecting, enabling on-demand dataset creation across 300+ millions of filtered records.
• Built a complete annotation ecosystem including task management, workflow orchestration, prioritization and quality controls — including AI model integration for automated blurring of sensitive/PII data, assisted ML annotation.
• Designed and implemented Video Access Manager, a secure video access service built on RBAC using Open Policy Agent combined with Authorisation service, enforcing fine-grained access control at scale.
• Contributed directly to improving Ring's ML models used across cameras and doorbells, with the platform serving as the primary data annotation pipeline for model training.
• Collaborated with ML engineers, product managers and data scientists to accelerate model development and experimentation.
• Mentored engineers and contributed to architecture decisions across multiple teams.