jonas@beck:~/portfolio
$

Jonas Beck.

I design and run the parts of the product nobody screenshots — APIs, queues, data pipelines and the on-call rotations that keep them honest. If it has a p99, I care about it.

0B+requests served / yr
0msflagship API p99
0%uptime, trailing 12mo
0TBdaily pipeline volume
##

systemctl list-units

# what I do
api-design.serviceREST & gRPC contracts people don't need a meeting to understand. Versioning that survives v3.active
distributed.serviceQueues, idempotency, backpressure, exactly-once-ish delivery with honest documentation about the "ish".active
data-pipeline.serviceStreaming and batch — CDC, Kafka, warehouse modelling. Data that arrives, reconciles, and is believed.active
reliability.serviceSLOs, error budgets, runbooks, chaos drills. Postmortems where nobody gets blamed and something gets fixed.active
##

GET /projects

# click a route
STREAM /projects/atlas-bus event backbone · 40k msg/s 200 OK

Company-wide event bus replacing a decade of cron jobs and CSV drops. Schema registry, dead-letter strategy, replay tooling. 40k messages/second at peak, with consumer lag alerts that actually mean something.

GoKafkaProtobuf40k msg/s11 teams onboard
GET /projects/ledgerd double-entry ledger · financial core 200 OK

Append-only double-entry ledger powering balances for 2M accounts. Strict serializability where it matters, eventual everywhere else. Survived an audit and a Black Friday in the same quarter.

RustPostgreSQLCQRS2M accountszero balance drift
POST /projects/shipyard internal deploy platform 201 Created

Golden-path deploy platform: one YAML file, canaries, automatic rollback on SLO burn. Took median deploy time from 40 minutes of ritual to 6 minutes of boredom.

KubernetesTerraformArgoCD120 servicesdeploys ×9/day
GET /projects/coldpath analytics pipeline · 3TB/day 200 OK

CDC out of the OLTP fleet into a lakehouse, with contract tests between producers and the warehouse. Analysts stopped asking "is this number right?" — the dashboards say when they aren't.

DebeziumSparkIcebergdbt3TB/day
##

git log --oneline career

# experience
v4.x · 2022 — now

Staff Engineer · Relay Systems

  • Own architecture for the payments core (ledgerd, atlas-bus)
  • Cut infra spend 31% while doubling traffic — mostly by deleting things
  • Run the architecture guild; write the RFCs people actually read
  • Removed: 14 services that should never have been services
v3.x · 2019 — 2022

Senior Backend Engineer · Hafen Logistics

  • Built tracking ingestion for 60k vehicles (MQTT → Kafka → Postgres)
  • Led the monolith strangler project — finished, which is the rare part
  • Introduced SLOs; pages went down 70%, sleep went up accordingly
v2.x · 2017 — 2019

Backend Engineer · Kassenwerk

  • POS sync engine for 3k stores with offline-first conflict resolution
  • First on-call rotation; learned the difference between alerts and noise
v1.x · 2015 — 2017

Junior Developer · Agentur Nord

  • PHP, MySQL, FTP deploys at 17:55 on a Friday — once
  • Deprecated: the habit of deploying at 17:55 on a Friday
##

htop — skills

# sorted by %CPU
processload%
go93
postgresql88
kafka85
kubernetes82
rust78
terraform76
observability70
css35