2026 Q2 — Available for projects

Production-grade AI & OSINT systems with measurable outcomes.

Founder at WinstonRed. I ship automation, OSINT platforms, and research engines with governance gates, test-first delivery, and clear operational handover.

Tests 1400+ Passing in CI
Apps 61+ Built in WinstonRedGuard
Governance Policy Gates ON Registry + quality checks
Snapshot --:-- Updated from current profile data

Need fast scope clarity? Book a 15-minute discovery call.

wrg — pulse & instinct

Yakuphan Yücel

I founded WinstonRed to create software that follows a simple philosophy: minimal dependencies, maximum reliability, everything tested.

My work spans OSINT platforms, multi-source research engines, sports analytics, AI-powered workflows, policy gate systems, and end-to-end application development. Every project ships with full test coverage and production-ready architecture.

PythonPowerShellFastAPISQLitePytestGitHub ActionsStreamlitRailwayOllamaRAGPlaywrightElectron
1400+
Tests Passing
61+
Apps Built
V1
Maturity
Baseline governance active

What I've Built

Featured
🛡️

WinstonRedGuard

The core monorepo powering all WinstonRed applications — 61+ apps, 1400+ tests, CI/CD green. A Python & PowerShell policy gate system with automated quality checks, governance rules, and a daily signal intelligence pipeline.

Problem: Many apps, fragmented standards, and high review overhead.
Solution: Registry alignment + automated policy gates + quality checks.
Outcome: 61+ apps, 1400+ tests, CI/CD consistently green.
PythonPowerShellGitHub ActionsMonorepo
Featured
🔬

Research Motor

Multi-source OSINT & research engine with 15 adapters (Reddit, GitHub, X, web scraper, repo audit, username OSINT, email intel, name lookup, and more). Watch mode, scheduled runs, HTML dashboard, and full CLI.

OSINT: Username scan across 68+ sites, email intelligence, name-based lookup.
Outcome: Repeatable intelligence reports and faster decision cycles.
Python180 testsOSINTCLIScheduler
Featured

Match Radar

Sports analytics pipeline that merges Forebet predictions, live odds, and statistical models into a unified signal. Kelly Criterion bet sizing, automated PDF reports, and a Streamlit dashboard.

Pipeline: Forebet → Odds API → Signal Merger → Kelly Criterion → Report.
Outcome: Automated daily match analysis with confidence-weighted sizing.
Python77 testsKelly CriterionStreamlit
📊

PulseBoard by WinstonRed

GitHub repo health dashboard with webhook auto-refresh, trend history, and tier-based pricing (Free/Pro/Agency). FastAPI + SQLite, GitHub OAuth, 5 async collectors, 0–100 Pulse Score. Live on Railway.

Problem: Repo health signals were scattered and hard to prioritize.
Solution: 5 async collectors + unified 0–100 scoring + one dashboard.
Outcome: A single operational view for health-based decisions.
FastAPI119 testsRailwaySaaS
🖥️

WRG Control Center

Electron-based desktop command center for the WinstonRedGuard ecosystem. OSINT dashboard, match radar controls, ecosystem overview, and real-time operational metrics.

FastAPI99 testsElectronOSINT
🎯

Paper Impact

Academic paper impact scorer. Fetches papers from arXiv, collects adoption signals across GitHub, social media, and news, then computes a 0–100 Impact Score with 5 weighted categories.

Problem: Research impact was hard to compare across disconnected signals.
Solution: Weighted scoring model combining arXiv, GitHub, social, and news signals.
Outcome: Repeatable 0–100 impact score for ranking and tracking papers.
Python156 testsarXiv APIZero deps
🤖

Agent Runtime

Lightweight agent execution engine with tool integration and chain-of-thought planning. Connected to Research Motor for a full artifact chain producing automated intelligence reports.

Python31 testsToolsAgents
🧪

Rule Lab

Policy rule authoring and testing workbench. 9 modules, 62 tests. Built from scratch as part of the WinstonRedGuard governance ecosystem.

Python62 tests9 modules
💬

Comment Copilot

LinkedIn comment generation assistant. Analyzer classifies topic/tone/intent, generator produces 5 styles, quality filter scores on a 100-point scale with spam detection.

Python89 testsStreamlitSQLite
🧠

Local AI Chat

Fully local AI chat powered by Ollama with RAG and ModelOps. SQLite vector store, FTS5 keyword search, model lifecycle management, and support for multiple models including Gemma and Qwen.

Python53 testsOllamaRAGModelOps

instinct-mcp

Self-learning pattern memory for AI coding agents. Observe repeated tool sequences, user preferences, and fix patterns — auto-promote to suggestions when confidence is high. Works with Claude Code, Cursor, Windsurf.

Problem: AI agents forget learned patterns between sessions.
Solution: Observe → track → promote pipeline with SQLite persistence and MCP integration.
Outcome: Agents build reusable instincts that improve over time.
PythonMCPPyPISQLite
📦

Affiliate Engine

Facebook affiliate post manager for home office niche. Product catalog, post generation, scheduling, tracking, and a Streamlit Cloud UI.

Python40 testsStreamlit

From Problem to Measurable Output

WinstonRedGuard

Problem: Many apps, fragmented standards, and review overhead.

Solution: Governance rules + automated policy gates + registry alignment.

Outcome: 61+ apps, 1400+ tests, CI/CD consistently green.

PulseBoard

Problem: Repo health signals were scattered and hard to prioritize.

Solution: FastAPI + SQLite dashboard with 5 async collectors and unified scoring.

Outcome: One operational view with actionable 0-100 pulse score.

Research Motor

Problem: Manual research cycles were slow and inconsistent.

Solution: 5-source adapter pipeline with watch mode and scheduled runs.

Outcome: Repeatable intelligence reports and faster decision cycles.

instinct-mcp

Problem: AI agents forget learned patterns between sessions.

Solution: Observe → track → promote pipeline with SQLite and MCP integration.

Outcome: Open-source on PyPI, reusable across Claude Code, Cursor, Windsurf.

How I Can Help

01

SaaS Development

End-to-end development of lightweight, test-driven SaaS products. From CLI tools to web dashboards, built with minimal dependencies and production-ready from day one.

  • Timeline: 2-6 weeks
  • Deliverables: architecture, MVP, CI/CD, handover docs
02

App Development

Custom Python applications with robust architecture. Automation pipelines, data processing tools, AI integrations, and full CI/CD setup with GitHub Actions.

  • Timeline: 1-4 weeks
  • Deliverables: tested codebase, CLI/API, deployment setup
03

Business Analytics

Multi-source data collection, automated scoring systems, and reporting dashboards. Turn raw data streams into actionable intelligence with scheduled pipelines.

  • Timeline: 1-3 weeks
  • Deliverables: scoring model, dashboard, scheduled reports
04

OSINT & Intelligence

Open-source intelligence platforms for username scanning, email investigation, and digital footprint analysis. Automated report generation with multi-source correlation.

  • Timeline: 1-3 weeks
  • Deliverables: OSINT toolkit, report pipeline, dashboard
05

MCP Server Development

Custom MCP server development for AI coding agents. Tool integration, memory pipelines, and multi-agent setups for Claude Code, Cursor, and Windsurf.

  • Timeline: 1-3 weeks
  • Deliverables: MCP server, tool definitions, PyPI package
06

Monorepo & Governance Setup

Multi-app monorepo architecture with automated governance pipelines, quality gates, release checks, and registry alignment. Battle-tested across 61+ apps.

  • Timeline: 2-4 weeks
  • Deliverables: governance rules, CI/CD pipeline, release gate
07

Data Pipeline & Scoring

Multi-source data collection pipelines with automated scoring algorithms and reporting dashboards. From raw signals to actionable 0-100 scores.

  • Timeline: 2-4 weeks
  • Deliverables: data pipeline, scoring model, dashboard
08

AI Agent Integration

Integrate AI agents into existing workflows. RAG pipelines, tool-use patterns, persistent memory, and local-first LLM setups with Ollama.

  • Timeline: 1-3 weeks
  • Deliverables: agent setup, RAG pipeline, memory store

Built for Reliability, Not Demos

Test-First Delivery

Core flows are covered with automated tests before handover to reduce regressions.

Low-Dependency Architecture

Lean stacks, explicit contracts, and operational simplicity for long-term maintainability.

Operational Handover

Delivery includes runbooks, CI/CD setup, and practical onboarding so teams can continue safely.

Let's Work Together

I'm open to freelance projects, collaborations, and interesting problems. Reach out through any of the channels below.

Start with a short brief

I usually reply within 24 hours.

Thanks, your message has been sent.