Looking for a job

Sergey Primosudov

@sergeyprimosudov

Message

Senior Go Backend Engineer at Connectly.ai who cut AI message-processing p95 latency from 237 to 113 ms.

Germany

Message

At Connectly.ai, I built Go backend components and a distributed processing pipeline for real-time messaging, supporting peak throughput of around 5K messages per second. I reduced p95 processing latency from approximately 237 to 113 ms through concurrency tuning, profiling, and benchmark-driven optimization.

Previously, I built backend services at Uber, Form3, and Grid Dynamics across mobility, real-time payments, and telecom. At Form3, my work supported payment processing workloads of around 19K RPS, and I reduced p99 latency from approximately 347 to 183 ms; at Grid Dynamics, services I built supported more than 21K concurrent WebSocket connections.

Experience

Work history, roles, and key accomplishments

Current

Senior Go Engineer

Current

Connectly.ai

Oct 2024 - Present (2 years 1 month)

Built high-performance Go backend components for real-time message ingestion and conversation processing, supporting peak throughput of around 5K messages per second. Reduced p95 processing latency from approximately 237 to 113 ms and improved stability under high concurrency.

Go Go Routines CHANNELS Context Synchronisation Pprof Benchmarking REST gRPC Protocol Buffers HTTP2 PostgreSQL Redis Kafka Test Containers

Senior Go Engineer

Uber

Feb 2022 - Sep 2024 (2 years 7 months)

Built high-throughput Go backend services for driver, trip, and account operations, supporting peak workloads of around 10K RPS. Reduced execution time of critical PostgreSQL queries by approximately 33% and increased test coverage from 68% to 82%.

Go gRPC REST PostgreSQL Redis Kafka Kubernetes Docker GoMock Test Containers CI CD DDD Saga Pattern

Go Engineer

Form3

Mar 2020 - Jan 2022 (1 year 10 months)

Built and maintained Go microservices for real-time payment processing and banking integrations, supporting peak workloads of around 19K RPS. Reduced p99 latency from approximately 347 to 183 ms and implemented idempotent financial transaction processing.

Go gRPC REST PostgreSQL Redis Kafka Idempotency Transactional Outboxes Pprof Query Optimization

Go Engineer

Grid Dynamics

Mar 2017 - Feb 2020 (2 years 11 months)

Built Go microservices supporting 21K+ concurrent connections and peak loads of around 15K RPS. Reduced p95 backend API latency by approximately 28% and improved production visibility with Prometheus, Grafana, and ELK.