AI & Software Engineer • Independent Contractor (UK Ltd)

I help companies optimize, scale, and modernize cloud-based systems, whether that’s high-throughput backends, streaming data platforms, or ML infrastructure. My focus: correctness, observability, and performance.

I’ve delivered solutions in e-commerce, telecom, and algorithmic trading, and helped clients scale workloads on Azure and AWS for significant cost and latency improvements.

• Backend & APIs: Elixir (Phoenix/Ash/Broadway), Python (FastAPI) - clean service boundaries, domain-driven design, reliable APIs.
• Data platforms: Kafka/RabbitMQ streaming, Dask/Pandas batch - backpressure handling, data contracts, high availability.
• Cloud & Infrastructure: Azure, AWS, Docker, CI/CD - system audits, scaling strategies, performance tuning, SRE practices.
• ML & RL engineering: PyTorch/Lightning - experiment pipelines, distributed training, model deployment and optimization.
• HPC & Performance: MPI/OpenMP/CUDA - parallel algorithms, profiling, large-scale compute.
• Trading & Feeds: resilient WebSocket ingestion at scale - exchange integrations, low-latency pipelines.

Engagements:

• 🚀 Project delivery (design & implementation)
• 🔍 Performance audits (bottlenecks, cost optimization)
• 🤝 Part-time retainers (ongoing support & scale-out strategy)
• 🧑‍💻 Interim engineering (helping teams bridge expertise gaps)

Availability: Remote (UK/EU-friendly) • On-site by arrangement

• nursery - Lightweight Elixir supervision library for isolating optional/costly subsystems.
• off_broadway_websocket - Resilient, backpressure-aware WebSocket ingestion for Elixir.
• async-websocket-pool - asyncio WebSocket pool for high-fanout market/data feeds.
• geminex - Elixir wrapper for the Gemini exchange REST API.

• Languages: Python/Cython, Elixir, C, Haskell
• Frameworks: Phoenix/Ash, FastAPI, Yesod
• Data & ML: Dask, Pandas, Broadway, PyTorch/Lightning, SQL/NoSQL
• Cloud & Infra: Azure, AWS, Docker, Kafka, RabbitMQ
• HPC & Parallelism: MPI, OpenMP, CUDA, OpenSHMEM

• ✅ Correct by design - explicit state, pure cores, property-based testing.
• 🛡 Resilient systems - retries, idempotency, observability, clear SLIs/SLOs.
• 📈 Performance-driven - profile before optimizing, tune for cost/performance balance.
• 🧩 Lean & modular - small surface area, clear contracts, minimal dependencies.

• 🌐 Website: michalpolit.com
• 💼 LinkedIn: linkedin.com/in/michal-p-ba1418198
• ✉️ Email: michal.polit@monadic.eu

Hiring? I specialize in helping teams scale cloud systems (Azure & AWS), optimize performance, and cut costs.
Let’s talk if you need a contractor who can deliver results fast.