● Built and operated the Go backend behind inDrive's real-time pricing — the price shown when a user opens the app and enters a destination. ● Owned services making pricing decisions across 14 AWS/EKS clusters, focusing on latency, correctness, and operational visibility. ● Refactored the MCP HTTP server, added metrics, dashboards, and alerts that made the pricing path easier to debug and support. ● Built backend flows for pricing experiments, releases, and switchbacks in partnership with A/B testing and ML teams. ● Tightened API/service boundaries and production runbooks for high-traffic user-facing flows.

● Migrated the flagship @Blum Telegram Mini App launcher bot from a TypeScript MVP to Go; profiled hot paths with pprof/Jaeger — new service sustained 200K concurrent users and 22M daily active users at peak. ● Designed and feature-led (team of 3) a Telegram Stars payment gateway, unblocking the bulk-send API critical to the token listing; earned 300K+ Stars (~$4K) in the first two months. ● Stabilized PostgreSQL access under listing load using external PgBouncer, mandatory pgx/v5 pooling, missing-index fixes, and lock fixes. ● Reworked high-traffic data access by moving hot OLTP paths to PostgreSQL and archiving cold data, reducing storage pressure and improving latency. ● Improved SDK-level logs/traces/metrics, led listing-related incident response, and saved ~400 GB in a sharded Redis cluster through low-risk Go optimizations.

● Built payment-antifraud scoring for SBP instant transfers (50%+ of all payment traffic) from scratch; mapped 11+ cross-team microservice integrations end to end. ● Redesigned the antifraud engine around a schemaless event/feature model, making API-partner integration faster and increasing test coverage via Go 1.22 generics. ● Fixed a hidden PgBouncer/PostgreSQL connection bug, raising daily credit-score recompute success for 2M+ users from 75% to 99.99%; added logging, tracing, and alerting to prevent regressions. ● Delivered ~50% of the work directly inside other teams' codebases, including PCI DSS scope, with careful review and rollout discipline.

● Built an algorithmic copy-trading backend for Base Chain (chain id 8453), evaluating routing APIs and integrating Zerion as the paid provider after technical discovery. ● Created an internal microservices API for blockchain event scanning (logs, tx.data) with Pub/Sub and polling interfaces, reducing RPC provider costs and simplifying backend integrations. ● Tech-led deployment of two production EVM chain forks (Vitru, Asset Chain): modified go-ethereum (geth), deployed boot nodes, and managed dev/stage/prod rollout across 24 servers with Ansible. ● Built AWS infrastructure as code in Terraform (EC2, S3, RDS, VPC, P2P over UDP) with CI linting and secure delivery checks. ● Delivered backend integrations across EVM, Solana, and TON, including Solana launchpad work and a Dedust SDK used by market-making systems.

● Planned and executed a zero-downtime split of a ~1 TB monolithic database into 24 shards, stabilizing catalog availability and reducing support load. ● Sped up Kafka consumers/producers 2.5x under strict resource quotas, making seller items purchasable immediately after creation. ● Designed and built a Go service with gRPC API and REST-style HTTP handlers for front-end and internal API partners. ● Built a product measurement automation pipeline handling 500K+ measurements per day across domains. ● Owned production reliability for catalog services: dashboards, alerts, postmortems, gradual migrations, and operational runbooks across ~11 services.