The Forward Deployment Engineer program: real AI engineering plus business English, taught live. Graduate with a deployed project, an English demo recording, and interview-ready skills. 12 sessions, ~15 hours total — Saturday 90 minutes + one weekday 60 minutes.
What a forward deployment engineer does. Real stack walkthrough: customer → API → model → data → monitoring. Live exercise: call an LLM API from Python with proper error handling and retries.
Standup format — "yesterday / today / blockers." Your first 60-second standup in English.
Production prompt design: system prompts, few-shot examples, structured JSON output. Build a small extraction tool. Intro to evals — how do you know your prompt is good?
Client writing: status-update emails. Rewrite a technical summary so a non-technical client understands it.
Embeddings, vector databases, retrieval + generation. Live build: a Q&A assistant over a sample document set.
Explaining trade-offs: present "RAG vs. fine-tuning" in English while the group plays skeptical client.
Containerize with Docker, wrap in a FastAPI endpoint, deploy to a free-tier cloud target. Environments, secrets, logs.
Incident communication: the client reports the assistant is down — diagnose in English, commit to a timeline.
Prompt-injection defenses, API key hygiene, rate limits, token budgets and caching. Add auth + guardrails to your deployment.
Proposal writing: draft a one-page "what we'll deliver and what it costs to run" in English.
Capstone: an end-to-end mini deployment for a fictional client brief, demoed live with peer code review.
Demo your capstone in English (5 min) and field questions. Feedback on clarity, not just correctness.
Engineers in LATAM (especially Mexico) with baseline Python. English can be weak — that's what the program fixes. 10+ seats per cohort. Graduates enter the S2HQ talent pipeline: deployed project, demo recording, and a direct line to US startups hiring.
