Python backends · Applied computer vision
I'm a computer engineering student focused on Python backend development and applied computer vision. I like digging into how a system works, then making its logic clear enough to test and improve.
Python · FastAPI · MediaPipe
Jump Jump has a browser workout interface and a typed HTTP API. It turns pose landmarks from squats, vertical jumps, and push-ups into exercise phases, rep counts, and rule-based feedback. I kept the exercise rules independent of MediaPipe so they can be tested with fixed inputs. Architecture · CI checks
Feedback accuracy against labeled real-world movement data has not been measured yet.
Python · aiogram · PostgreSQL · Redis
SubMate is a Telegram bot for finding movie and TV subtitles. It uses TMDb to identify a title or episode, searches OpenSubtitles, and validates the SRT file before sending it. I kept search and delivery outside the Telegram handlers; PostgreSQL and Redis are optional adapters. Architecture · Quality checks
Automated tests use fake providers. Live movie and episode delivery still need to be checked with the real services (integration checklist).
I'm working on a read-only Dagster workflow for GitHub issues. I'll add a case study after I verify a live run.