raxhacks // ai-infra-swe

Raxhacks AI Infra SWE Learning Roadmap

A 4-week, hands-on path from Linux and containers to production inference infrastructure. Check off each day, log your notes, and ship the weekly artifact as you go.

ARTIFACT

Dockerized microservice with scripts

url

D1Set up the terminal, shell and dotfiles
Notes for day 1
D2Linux filesystem, processes and permissions
Notes for day 2
D3Networking fundamentals: TCP/IP, DNS and HTTP
Notes for day 3
D4Bash scripting and automation
Notes for day 4
D5Docker images, containers and Dockerfiles
Notes for day 5
D6Docker Compose and multi-container services
Notes for day 6
D7Resource limits, logs and debugging containers
Notes for day 7

ARTIFACT

Three-service app on k8s via Helm

url

D1Kubernetes architecture and core objects
Notes for day 1
D2Deployments, pods and services
Notes for day 2
D3ConfigMaps, secrets and persistent volumes
Notes for day 3
D4Ingress, service mesh and cluster networking
Notes for day 4
D5Helm chart fundamentals
Notes for day 5
D6Observability: logs, metrics and traces
Notes for day 6
D7Kubernetes troubleshooting drills
Notes for day 7

ARTIFACT

Containerized inference service with API

url

D1Model formats: ONNX, safetensors, GGUF
Notes for day 1
D2Inference servers: vLLM, TGI, Triton
Notes for day 2
D3GPU scheduling and batch inference
Notes for day 3
D4Designing REST and gRPC inference APIs
Notes for day 4
D5Autoscaling and queues for inference
Notes for day 5
D6Quantization, caching and latency tuning
Notes for day 6
D7Load-testing inference endpoints
Notes for day 7

ARTIFACT

End-to-end AI infrastructure capstone

url

D1Feature stores and data pipelines
Notes for day 1
D2Model lifecycle from training to production
Notes for day 2
D3CI/CD for ML pipelines
Notes for day 3
D4Monitoring drift, latency and SLOs
Notes for day 4
D5Cost optimization and capacity planning
Notes for day 5
D6Security, multi-tenancy and governance
Notes for day 6
D7Capstone: run an AI platform end to end
Notes for day 7