Paweł Oljasz
|
I architect and secure complex systems - from GPU clusters and real-time data pipelines to enterprise SOC platforms and IoT networks. Leading a team of senior specialists across infrastructure, security, and software delivery. Over 15 years building things that run in production and stay up.
What I Work With
Infrastructure
Distributed clusters, containers, observability at scale
AI & GPU
CUDA, LLM deployment, distributed inference
Cybersecurity
Threat detection, LLM safety, compliance
Blockchain
Smart contracts, DeFi, node infrastructure
Embedded & IoT
Microcontrollers, edge computing, firmware
Home Automation
Autonomous systems, energy, sensor fusion
Leadership
High-performing teams, technical strategy
Robotics
Computer vision, path planning, real-time control
Networking
SD-WAN, VPN, firewall, load balancing, DNS
SIEM & SOC
Splunk, ELK, threat hunting, incident response
Big Data
Data lakes, Spark, Kafka, real-time pipelines
Cloud
AWS, Azure, GCP, multi-cloud, migration
Recent Writing
All postsSilent Leadership: A Manifest
Why I invest in people over processes, build teams that understand products deeper than the vendors themselves, and only open support cases when we find real flaws.
Modbus RTU Over USB: Five Lessons From a Year of Polling Solar Inverters
USB-RS485 adapter quirks, pymodbus CPU bugs, silent data corruption, and the wiring trap that cost me a full day. Hard-won lessons from running Modbus RTU on a Raspberry Pi 24/7.
Two Docker One Bug
A stale DNS forwarder in an 8-month-old container. A missing env var from a forgotten Compose overlay. Two silent failures, same evening, both hiding in plain sight.
Deploying a Next.js Site to Cloudflare Workers - The Full Walkthrough
End-to-end guide to deploying a Next.js site on Cloudflare Workers with a custom domain. The gotchas nobody warns you about.
Adding an RTX 5090 GPU Server to a Mac Mini Cluster
A Ryzen 9 9950X3D + RTX 5090 build running Ubuntu, Ollama, and ComfyUI. Custom ISO, driver gotchas, and benchmarks.
Building an LLM Security Pipeline - Guard, Classify, Log Everything
LLM Guard for input scanning, Llama Guard 3 for output classification, Langfuse for observability. A complete security pipeline for self-hosted LLMs.