Forward Deployed Engineer, requirement 12 of 12
Client-facing delivery
I turn engineering work into updates that client architects, security leads and business owners can each act on.
The requirementWork with client teams, architects, security, infrastructure and engineering staff, and communicate with each in their own terms.
How I deliver it at a customer
- Map the stakeholdersI list every client architect, security owner and infrastructure lead in the workspace CRM and contacts app. Each contact record names the one decision that person owns, so every update has a named reader.
- Index their languageTheir runbooks, architecture notes and policies go into the retrieval service through its document ingest endpoint. Access is checked per persona before retrieval, and queries come back in the client's own system names.
- Draft the updateStatus drafts run through the multi-provider model gateway, which holds the provider keys, applies rate limits and logs every call. Tokens are scoped to an allowed list of apps, and the log shows which model produced a draft.
- Cut the jargonI read the draft and replace terms like KEDA or CloudNativePG with wording a business owner can act on. The demo on this page runs that same pass on Cloudflare Workers AI.
- Sign it myselfApproval gates and audit on agent tools are Early access, so I do not rely on them for client mail. I read and sign every outgoing update myself, and nothing agent-written leaves unsigned.
- Send and file itThe update goes out from the workspace mail app, and the review is booked in the calendar. On the next mail sync, the thread is logged as an activity against the matching lead.
Tools I use
| Tool | What it does here | Status |
|---|---|---|
| Multi-provider model gateway | Routes every drafting call, holds the provider keys and rate limits, and records each call for later review. | Available |
| Document retrieval service | Returns the client's own runbooks and decisions, so an update uses their terminology instead of mine. | Available |
| Workspace apps | CRM, contacts, mail and calendar hold the stakeholder list, send the update, book the review and log the thread. | Available |
| Approval gates and audit | Would put a named person in front of anything an agent sends outside the engineering team. | Early access |
| Cloudflare Workers AI | Runs this page's rewrite and jargon list on Llama and bge models; it is not a customer deployment. | Demo only |
Available means shipped and in use. Early access means built but not generally available. Roadmap means designed, not built. Demo only means it powers this public demo, not customer deployments.
This demo runs on Cloudflare Workers AI and is not a customer deployment. Paste rough engineering notes and it returns a status update written for a business reader. It also lists each jargon term with a plain replacement.
Open the live demoThe other answers
LLM applications in productionEvery model call goes through one gateway, guardrails run on both sides, and each task gets a measured model choice.Prompt engineeringI write the output shape first, version the prompt, budget its tokens, and compare model replies before shipping.Retrieval-augmented generationGrounded, cited answers over the customer's own documents, with access rules applied before retrieval.AI agents and orchestrationI build agents that call your systems as registered tools, and the workflows that trigger and sequence them.Request classification and routingI build automation that classifies each inbound request, routes it by role, and holds risky changes for human approval.AI workflow orchestrationI design the steps, contracts, and data flows, then ship them as running workflows on customer systems.API and integration designI publish the API contract first, gate it with keys and scopes, then build integrations against it.Hybrid enterprise integrationI connect AI services to your identity, applications and data, then run them where your policy allows.Prototype to productionI take a prototype through generated test suites and merge review gates to a zero-downtime deploy that rolls back on failure.Approvals and exception handlingI put a person between the agent and the change, with gated tools, a pending queue, expiry, and a record.Evals and output validationI score answers against a versioned rubric, trace failing calls to the integration that broke, and tune latency per call.
Gus IT LLC. Contact gus@gusit.de. Claude is a product of Anthropic PBC. Cloudflare and Workers AI are trademarks of Cloudflare, Inc. Llama is a trademark of Meta Platforms, Inc. Other names are trademarks of their respective owners. Gus IT LLC is an independent engineering firm and is not affiliated with or endorsed by these companies.