Full Stack Engineer - Agentic AI | Python | Typescript | PHP | K8 | Terraform

You’ll architect, code, and operate the services that power our AI\-first platform. From Django APIs and raw\-PHP back\-ends to virsh\-provisioned VMs and H100/H200 GPU clusters, you’ll own tough problems end\-to\-end—and ship rock\-solid solutions that stay up when you’re offline. What You’ll Do: •  Backend \& Full\-Stack Development   ◦  Design, implement, and review Django services and REST endpoints (Python 3\.9\+).   ◦  Maintain and extend legacy/raw PHP modules—no frameworks, just clean code.   ◦  Build or tweak lightweight front\-end dashboards when needed using HTMX and vanilla JS.   •  DevOps \& Infrastructure   ◦  Automate CI/CD (GitHub Actions, GitLab CI, or similar).   ◦  Provision and manage VMs \& virtual networks via virsh/libvirt; codify it with Ansible or Terraform.   ◦  Instrument, log, and alert with Prometheus/Grafana, ELK, etc.   ◦  Kubernetes (nice\-to\-have): GPU\-aware scheduling, Helm charts, and cluster debugging.   •  GPU \& AI Workloads   ◦  Configure, profile, and optimize NVIDIA H100/H200 servers (drivers, MIG, NCCL, NVLink topologies).   ◦  Serve and fine\-tune models with Triton, TensorRT, vLLM, or Hugging Face stacks.   •  Prompt Engineering   ◦  Craft, test, and iterate high\-impact prompts for LLM features; collaborate with data scientists to boost output quality.   •  Problem Solving \& Support   ◦  Triage production incidents, drive root\-cause analyses, and write blameless post\-mortems.   ◦  Partner with product \& customer success to deliver quick, durable fixes. Must\-Have Qualifications •Python \& Django I◦ diomatic, test\-driven code; class\-based views, DRF, async tasks, ORM tuning. •Raw PHP ◦ Maintain and create uncluttered PHP 8\.x scripts/modules (no Symfony, no Laravel). •DevOps ◦ Docker/Podman, CI/CD pipelines, infra\-as\-code, secrets management. •VM orchestration ◦ Daily hands\-on with virsh/libvirt (snapshots, live migration, virt\-network). •GPU expertise ◦ Deploy, monitor, and tune H100/H200 clusters; CUDA, NCCL, MIG partitioning. •Debugging chops ◦ gdb/LLDB, strace, tcpdump—whatever cracks the bug. •Communication ◦ Clear, concise English; you document systems and decisions. •Timezone overlap ◦ Comfortable covering core hours that touch both PST (UTC\-8\) and EST (UTC\-5\). Nice\-to\-Haves •  Kubernetes (GPU scheduling, Helm, Karpenter).   •  HTMX wizardry for dynamic server\-driven UIs.   •  Distributed tracing (OpenTelemetry, Jaeger).   •  Payments or comms integrations (Stripe, Twilio, etc.). Soft Skills We Value •  Relentless curiosity — you dig until you truly understand.   •  Bias for action — you turn ideas into shipped code quickly.   •  Ownership — you build systems you’re proud to support (even on\-call).   •  Team spirit — you mentor, pair program, and celebrate wins together. Benefits •  Competitive salary •  Fully remote; choose your workspace (laptop provided).   * •  Annual budget for conferences, training, and hardware.