Forward AI Deployment Engineer (FDE)

SigmaSoftware2

1h 30m ago

On-site

Full Time

Poland

Embed directly with the client's operational team to find the workflow that's actually costing them time, quality, or money before recommending any tooling.
Own the redesign end to end: rework the decision flow and handoffs around AI, then take the resulting system into production yourself.
Translate a vague, cross-functional business problem into a concrete technical plan the client's team can actually execute against.
Prototype fast enough to show the client what's possible, then build only the parts of that prototype the redesigned workflow actually needs.
Where the redesign calls for it, build the agentic systems, RAG pipelines, or custom data pipelines that make it real never a deliverable in itself, always because the workflow needs it running in production.
Manage stakeholder alignment across the client's org, escalating early and requesting support before a blocker becomes a delay.
Build up the client's own team alongside the system pair with their people, explain the reasoning behind each decision, and leave them able to run and improve what you built after you've moved to the next engagement.
Build the feedback loop from day one telemetry, evaluation pipelines, human-in-the-loop review — so the system keeps improving after you've moved on.

Proven experience taking AI systems from prototype to production inside a real operating environment, not just in a lab or sandbox.
You treat LLM/RAG architectures, agentic systems, cloud platforms, and data pipelines as tools for redesigning how work gets done, not as the deliverable — and you're strong and current enough with all of them to actually use them that way.
Comfort mapping decision logic that's partly undocumented, and judgment about where automation should stop and human accountability should start.
Excellent communication and stakeholder management — able to advise a client directly and escalate without being asked.

llmai

🇬🇧 English

SigmaSoftware2

1h 30m ago

On-site

Full Time

Poland

Embed directly with the client's operational team to find the workflow that's actually costing them time, quality, or money before recommending any tooling.
Own the redesign end to end: rework the decision flow and handoffs around AI, then take the resulting system into production yourself.
Translate a vague, cross-functional business problem into a concrete technical plan the client's team can actually execute against.
Prototype fast enough to show the client what's possible, then build only the parts of that prototype the redesigned workflow actually needs.
Where the redesign calls for it, build the agentic systems, RAG pipelines, or custom data pipelines that make it real never a deliverable in itself, always because the workflow needs it running in production.
Manage stakeholder alignment across the client's org, escalating early and requesting support before a blocker becomes a delay.
Build up the client's own team alongside the system pair with their people, explain the reasoning behind each decision, and leave them able to run and improve what you built after you've moved to the next engagement.
Build the feedback loop from day one telemetry, evaluation pipelines, human-in-the-loop review — so the system keeps improving after you've moved on.

Proven experience taking AI systems from prototype to production inside a real operating environment, not just in a lab or sandbox.
You treat LLM/RAG architectures, agentic systems, cloud platforms, and data pipelines as tools for redesigning how work gets done, not as the deliverable — and you're strong and current enough with all of them to actually use them that way.
Comfort mapping decision logic that's partly undocumented, and judgment about where automation should stop and human accountability should start.
Excellent communication and stakeholder management — able to advise a client directly and escalate without being asked.

llmai

🇬🇧 English