Skip to content
View KapaSique's full-sized avatar
🪼
china exchange
🪼
china exchange

Block or report KapaSique

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
KapaSique/README.md

Artem Svinoboev · KapaSique

CS student → ML/CV & full-stack engineer.
I architect systems and orchestrate AI agents to ship them — fast, but with my hands on the wheel.


// about

  • 3rd-year Informatics & Computer Engineering student at NEFU (Yakutsk), grad ~2027.
  • Currently on exchange at Jiamusi University, China — learning Mandarin on the ground.
  • Heading toward a Master's in AI / CS in Shenzhen / Guangzhou. The neon-and-silicon city stack is the goal, not the backdrop.
  • ~4 years building web; expanding into mobile and computer vision.

// how I work

I'm not a "write every line by hand" dev and I don't pretend to be. I design the architecture, write the specs, and direct AI agents to execute — then I review, integrate, and own the result. The skill isn't typing; it's knowing what to build, how it fits together, and where the agents are wrong.

spec & architecture   → me
heavy lifting / code   → Claude (Opus) via Claude Code
planning / TZ          → GPT / Codex
high-volume grind      → DeepSeek (API)
agentic dev env        → Cursor + Claude Code
review & integration   → me

// selected work

  • 🏎️ f1-pitstop-prediction-s6e5 — Kaggle Playground S6E5. GBDT + RealMLP blend, OOF-validated. Private AUC 0.9545 (~top 7%).
  • 👷 CV worker-monitoring system on Raspberry Pi 5 — real-time computer vision on edge hardware. Funded by an FSI UMNIK grant (~500K ₽).
  • 🧪 Selected for SMILES-2026 (Summer School in Machine Learning).

// stack

// currently

  • 🇨🇳 Studying Mandarin (HSK path) + prepping IELTS/TOEFL for grad applications.
  • 🔭 Building ML/CV side projects and sharpening the agent-driven workflow.
  • 🏃 Off-screen: running toward a half-marathon, lifting, and a soft spot for Mercedes.

// reach

Ship real systems. Stay honest about how. 深圳, here I come.

Pinned Loading

  1. f1-pitstop-prediction-s6e5 f1-pitstop-prediction-s6e5 Public

    🏎️ Kaggle Playground S6E5 — F1 Pit Stop prediction. GBDT + RealMLP blend, OOF-validated. Private AUC 0.9545 (top ~7%).

    Python

  2. stellar-class-prediction-s6e6 stellar-class-prediction-s6e6 Public

    🌌 Kaggle S6E6 — Stellar Class (GALAXY/QSO/STAR). GBDT ensemble + RealMLP + my own LR-stacker (with P100 torch fix). Balanced Accuracy 0.9711, top ~8%.

    Python

  3. SMILES-2026 SMILES-2026 Public

    Zero-order fine-tuning of ResNet18 on CIFAR-100 — 49.68% top-1 with no gradient computations (SMILES-2026 ML summer school).

    Python

  4. second-look-triage second-look-triage Public

    Second Look — a clinically-grounded triage safety-net & audit system (Triagegeist hackathon): calibrated ESI + vitals-independent red-flag NLP + honest informative-missingness & fairness audit + li…

    Python