specialone0007 (Furkan Reha) · GitHub
Skip to content
View specialone0007's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report specialone0007

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.

Content in all repositories owned by your account will be closed.
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
specialone0007/README.md

Furkan Reha Tutas

Software engineer in Istanbul. I build products end to end in Python and TypeScript, and for the last two years mostly products with LLM agents inside them: the billing, the evals and the deterministic checks that turn an agent demo into something people pay for.

Currently at Swarm Network, remote.

Live products

What it is My part
Eclipse FM AI-native music platform. People live-code music while AI agents compose alongside them. Built end to end: React, Express, Python. Multi-agent composition system (a manager agent over four instrument specialists), credit and subscription billing, data model.
43 AI prediction platform. Agents reason over each question and publish their forecasts in public. Built end to end: Next.js, TypeScript, Python. Web app, ingest bot, credit economy, backtesting harness, and the settlement pipeline with 13 code gates between the LLM judge and the database write.
Vault SDK Secrets manager for AI agents on Sui. Developer experience: the documentation site and the npm releases of @swarmnetwork/vault-sdk.
4Fans Football fan community with live scores for 125 competitions. Built solo over 14 months and run it: FastAPI, PostgreSQL, 160+ endpoints, 3,200+ tests.

Also at Swarm: the draw engine, signed reward-claim flow and operations console behind the Agent License Sale (10,000+ NFT licenses, $10M+ on Sui), and swarm-skills, seven review and audit skills for AI coding agents, where a fresh critic scores every change and the next critic re-checks each fix. It runs as the merge gate on our repositories.

On-chain

  • Wrote the Hacken-audited Solidity for AgentXYZ on Avalanche: the TRADER token and its vesting contract, Hardhat and ethers tooling, through audit to mainnet.
  • Audited and hardened an NFT rewards protocol paying USDC to holders: upgradeable reward vault, EIP-712 signed receipts, payout ledger, claim verification.
  • Designed token models for 30+ launches at CodeCraft Group and Seedify.

Earlier work, public here

Rewritten in 2026 as tested packages with CI, each with a short report:

Contact

LinkedIn · tutasfurkanreha@gmail.com

Popular repositories Loading

  1. vertexCover vertexCover Public

    Greedy 2-approximation for Minimum Vertex Cover in C++17: exact solver for ground truth, fuzz tests, running-time and quality benchmarks, CI.

    C++

  2. detectingPhishingWebsites detectingPhishingWebsites Public

    Phishing website detection on the UCI Phishing Websites dataset: tested Python package, 8 classifiers, honest hold-out evaluation, reproducible figures.

    Python

  3. labelMicroTopicRelationships labelMicroTopicRelationships Public

    Labelling prerequisite relationships between micro-topics with as few expert questions as possible: simulated oracle, six query strategies, accuracy-vs-questions benchmark. Graduation project, rewr…

    Python

  4. footballDataVisualization footballDataVisualization Public

    Match-level football visualisations from StatsBomb event data: xG race, shot map, average positions, heatmaps, shot sequences. Python package + CLI, tested.

    Python

  5. transferCapabilitiesOfBlendganByLeveragingInterfacegan transferCapabilitiesOfBlendganByLeveragingInterfacegan Public

    Semantic face editing with InterFaceGAN followed by artistic stylisation with BlendGAN; comparison of three GAN-inversion approaches. Report + result figures (2022 course project).

  6. wsnGameTheory wsnGameTheory Public

    Game-theoretic jamming simulation for wireless sensor networks (Mao, Zhu & Wei 2013): vectorised NumPy game, best-response dynamic, parameter sweeps, tests, CI.

    Python