class AliPourrahim:
name = "Ali Pourrahim"
location = "Finland 🇫🇮"
hireable = True
languages = ["Python", "Go", "TypeScript", "SQL", "Bash"]
backend = ["FastAPI", "Go stdlib", "SQLAlchemy", "asyncpg", "gRPC"]
cloud = ["Azure", "Docker", "Kubernetes", "Terraform"]
databases = ["PostgreSQL", "Redis", "MongoDB"]
ai_ml = ["LLMs", "sentence-transformers", "LangChain", "conformal prediction",
"OpenAI", "scikit-learn", "Hugging Face"]
observability = ["Prometheus", "OpenTelemetry", "Grafana", "Sentry"]
messaging = ["Kafka", "Azure Service Bus", "WebSocket"]
currently = "Building AI-powered developer tooling and production ML systems"
def say_hi(self):
print("Building things that actually work in production.")
| Project | What it does | Stack |
|---|---|---|
| bitcoin-mcp | 49-tool Bitcoin MCP server for AI agents — fees, mempool, blocks, mining, price. Zero config | Python, MCP, FastAPI |
| AgenticAIPoweredLinuxSystemManagement | 6-layer agentic AI pipeline: NLP intent parsing, risk policy gating, safe execution, audit log | Python, LLM, SQLite |
| codebase_intelligence | AI-powered semantic codebase search, dependency analysis, automated documentation | Python, LLM, FastAPI |
| legalens | Semantic diff engine for legal documents — clause-level meaning shifts and risk scoring | Python, FastAPI, sentence-transformers |
| culture-identifier | NLP personality analyzer matching writing style to iconic thinkers via semantic embeddings | Python, FastAPI, Transformers |
| HAIIP | EU AI Act compliant industrial AI platform for Nordic SMEs | Python, Jupyter, FastAPI |
| Project | What it does | Stack |
|---|---|---|
| cloudcostguard | Scans AWS and Azure for idle VMs, unattached disks — reports exact dollar savings | Go, Prometheus, Docker |
| cloud-calibration | Cloud-native calibration measurement platform with Azure Service Bus and REST API | Go, Azure, PostgreSQL |
| observability-platform | Distributed tracing, metrics aggregation, real-time alerting | Go, OpenTelemetry, Prometheus |
| multi-tenancy-kubernet | Multi-tenant SaaS on Kubernetes: namespace isolation, Helm automation, RBAC | Shell, Kubernetes, Helm |
| distributedJobscheduler | Redis-backed distributed job scheduler with atomic locking and fault-tolerant execution | Go, Redis, gRPC |
| Project | What it does | Stack |
|---|---|---|
| mm-live | Event-driven market making: Binance WebSocket, Kalman fair value, Avellaneda-Stoikov quoting | Python, asyncio, Binance WS |
| backtest-audit | Statistical overfitting audit for trading backtests — Deflated Sharpe Ratio, Monte Carlo permutation | Python, scipy, pandas |
| Fingrid-dashboard | Real-time Finnish electricity market monitoring via Fingrid Open Data API | Python, FastAPI, PostgreSQL |
| streaming-data-pipeline | Real-time pipeline with layered anomaly detection (Welford + EWMA) and Kafka | Go, Kafka, WebSocket |
| logforge | Production log analytics: Kafka ingestion, PostgreSQL storage, real-time alerting | Python, FastAPI, Kafka |
| Project | What it does | Stack |
|---|---|---|
| Finnish-nlp-2.0 | World-class Finnish NLP: 30+ endpoints, 99 tests, 100% coverage — all 15 grammatical cases | Python, Voikko, FastAPI |
| Project | What it does | Stack |
|---|---|---|
| PipelineGuard | CI/CD security: automated quality gates, vulnerability scanning, deployment guardrails | Python, FastAPI, PostgreSQL |
| Student-Retention-Prediction | Production ML for dropout risk prediction — 137 tests, 100% coverage | Python, scikit-learn, FastAPI |
| Project | What it does | Stack |
|---|---|---|
| cascade-conformal | Provably valid prediction sets for K-stage ML pipelines. Five theorems, production library, O(log B) streaming updates. | Python, NumPy, SciPy |
| conformal-risk | Conformal risk control for general loss functions in production ML | Python, scikit-learn |
"The best infrastructure is the one no one notices — because it always works."