Igor Savinkin

Contributed backend APIs for an RTSP video ML service (FastAPI health, predict, SSE); integrated pre-processing and inference pipeline; containerized runtimes; implemented CI/CD automation with GitHub Actions to push images to AWS ECR and deploy to ECS Fargate; added CloudWatch logging/metricks; integrated AI APIs for data enrichment and intelligent processing.

Specialize in rapid prototyping and “vibe-first” feature development, quickly transforming ideas into functional, testable product experiences and MVPs. Validate hypotheses through fast iteration cycles, building lightweight prototypes to assess user “feel” feasibility and technical potential. Deploy solutions locally (Windows + Python) and on VPS environments (Linux). Actively use Cursor as a primary code-generation and integration platform to accelerate delivery and maintain development efficiency.

Built backend services to collect, normalize, and store large datasets from web sources; designed ETL-like pipelines converting unstructured web data into structured SQL datasets; developed Python-based document processing for automated proofreading; improved backend performance via asynchronous processing and multi-threading; collaborated with stakeholders to translate data needs into technical solutions; key project: Stream Video ML Backend Service (AWS, FastAPI, Docker); exposed REST and SSE APIs; deployed containerized services to AWS ECS Fargate with CI/CD; implemented observability using CloudWatch; built data capture pipelines for model retraining.