Blog
1919+ articles — Page 7 of 80
Snow Day for Your Infrastructure: Are You Prepared?
No bad weather, only infrastructure that isn't prepared. My snow-day take on traffic spikes, black swans and a readiness checklist for Kubernetes teams.
12 min read
A2A Python SDK Tutorial: Server, Client and Streaming
A2A Python SDK tutorial with a2a-sdk 1.2: AgentExecutor, Agent Card, Starlette server, client, tasks, streaming, input-required, cancel and an MCP tool.
15 min read
Builders & Brews Amsterdam: Nebius x NVIDIA AI Hackathon
AAIF Community Amsterdam's Builders & Brews: Hack Edition: the hackathon rules, Token Factory, Tavily, Kilo Code and a fast agent recipe before 30 October.
15 min read
AI Native Netherlands at Elastic: Agents, MCP, Workflows
AI Native Netherlands at Elastic Amsterdam: ANWB on forecasting roadside assistance demand, then Elastic on moving from RAG to agents, MCP and workflows.
13 min read
AI Tech Summit Skopje 2026: From AI Demo to Production
My keynote at AI Tech Summit Filip Avramchev 2026 in Skopje, the sessions I photographed, and a practical checklist for taking AI from demo to production.
12 min read
Amsterdam Tech Meetups, January 2026: A Roundup
Five January 2026 evenings: a YC founder panel at Miro, Polarsteps product principles, BitBash Jurassic Edition, a Xebia book launch and a pitch night.
7 min read
Ansible Conjur Tutorial: Secrets with cyberark.conjur
A tested Ansible Conjur tutorial: run Conjur OSS in Docker, load a host policy, fetch a password with the cyberark.conjur lookup, hide it from logs, rotate it.
13 min read
ArgoCD Pull Request Generator: Per-PR Preview Environments
Build per-PR preview environments with the ArgoCD pull request generator and Gitea on kind: one namespace per PR, SHA image tags, cleanup on close, quotas.
12 min read
The 6-Step Sales Funnel for B2B AI Products
How to adapt the 6-step B2B sales funnel for AI product companies — with ad angles, call sequences, and revenue metrics that convert.
11 min read
Bazel Multiple Python Versions with rules_python
Bazel multiple Python versions in one monorepo: two rules_python toolchains, per-version pip locks, python_version pins, cquery checks and real pitfalls.
11 min read
Bazel Remote Cache with bazel-remote: Setup and Debugging
A Bazel remote cache with bazel-remote: action cache vs CAS, --remote_cache over gRPC, cold vs warm numbers, read-only CI and the cache-key traps I hit.
12 min readBuildbarn on Kubernetes: Bazel Remote Execution Lab
Buildbarn on Kubernetes with the official bb-deployments manifests on kind, Bazel actions running remotely, cache hits, platform mismatches and worker scaling.
15 min read
Community Days Milan 2015: Microsoft IoT Then and Now
Slides I photographed at Community Days 2015 near Milan, from Event Hubs to Windows Embedded and Intel Galileo, and what each one became by 2026.
9 min read
CTO Network State of Agents 2026 at Databricks Amsterdam
A Builders CTO Network breakfast at Databricks Amsterdam: four people who ship agents, five trend slides and a closing look at the software factory.
11 min read
Datatrove: Build an LLM Data Curation Pipeline, Measured
A datatrove LLM data curation pipeline on a messy test corpus: licence filtering, Trafilatura, language ID, Gopher/C4, MinHash dedup and PII, with counts.
15 min read
Debug Bazel Build Failures: BEP, Exec Logs, Sandbox
How I debug a Bazel build: four failures reproduced on purpose and pinned down with the sandbox, aquery, toolchain debug, exec log diffs, BEP and profiles.
15 min read
Docling Tutorial: PDF to Markdown, Tables, OCR and RAG
A Docling tutorial tested on four real PDFs: Markdown and JSON export, OCR on a 1933 scan, tables to pandas, chunking and a local Chroma RAG with Ollama.
17 min read
Elastic Agent Builder MCP: Tools and Workflows, No LLM
Run Elastic 9.5 locally, build ES|QL and workflow tools, write an Elastic Workflows YAML file and call it all over the Agent Builder MCP endpoint, no LLM.
14 min read
Hey Europe, Wake Up on AI: A Call From Paris
In January I filmed a call to action in front of the Petit Palais in Paris: Europe built palaces like this, so it can build AI infrastructure and models too.
7 min readGranite Guardian on Ollama: Local LLM Guardrails
Run IBM Granite Guardian 4.1 on Ollama as a local guard model: screen prompts, check RAG groundedness and tool calls, with a Python gate and measured results.
14 min read
kube-proxy iptables Mode: How Kubernetes Services Work
kube-proxy iptables mode walked chain by chain on kind: KUBE-SERVICES, KUBE-SVC, KUBE-SEP, random load balancing, NodePort, conntrack, then nftables mode.
14 min read
MLOps Community Amsterdam: AI Sovereignty and GPT-NL
The MLOps Community Amsterdam sovereignty special at UvA: a policy panel on Europe's compute gap and TNO's talk on training GPT-NL, a Dutch LLM, from scratch.
11 min read
Neo4j Vector Index for GraphRAG: Search, Then Traverse
Create a Neo4j vector index on chunk nodes, query it with SEARCH, expand the hit into related entities in the same Cypher query, and add full-text hybrid.
14 min read
OpenTelemetry Baggage in Python Without Leaking It
OpenTelemetry baggage in Python: the W3C header, reading and writing it, BaggageSpanProcessor, and a tested lab that stops it leaking to third-party APIs.
11 min read