Ten years of shipping backends, data platforms, and ML/AI that survive real load and real users. I like owning a service from the architecture doc to the 2am page. Here's what I do — and how to get my CV in one click.
FOR COMPANIES
What I do
-
Backends & distributed systems — typed async Python (FastAPI, Django, Celery), API/SDK design, sub-second latency.
-
Data platforms & pipelines — GCP, AWS, ClickHouse, dbt, Airflow. Messy sources → trustworthy tables.
-
Reliability & security — monitoring/SLOs (Prometheus/Grafana/Loki), enterprise auth (SSO/SAML, FIDO2/passkeys), CVE remediation, zero-downtime ops.
-
MLOps & LLM systems — inference, monitoring, cost control, RAG & agents that survive real users.
-
Making things cheaper — my favourite invoice line is the one that goes down.
Selected results (proof)
-
Doubled peak throughput to 100 RPS and held it steady.
-
Deleted 70M+ live rows and freed 140 GB on a production database — zero downtime.
-
Remediated 10 critical/high CVEs across production images with zero-downtime redeploys.
-
Cut a firefighting backlog 40%, landed a 4.85% change-failure rate.
