Forward Deployed Specialist

Elios AI; San Jose (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

Elios AI is seeking a Forward Deployed Specialist to work inside enterprise accounts, turning AI capability into production software that actually runs. You will split time between understanding client problems and building the integration, ensuring the solution stands up in production.

You will own the customer relationship and drive technical design, deploy LLM-powered apps, and guide pilots from workshops to production handover while upholding security and governance.

Qualifications

• 5+ years building and shipping production software.
• Comfort deploying and operating on cloud platforms (AWS/GCP/Azure).
• Experience integrating with enterprise systems and auth/services.

Responsibilities

• Embed with enterprise customers to map workflows and constraints.
• Build and ship LLM-powered applications and agentic workflows.
• Wire into customer systems including legacy APIs and on-prem data stores.

Skills

Python

APIs

Cloud platforms

Tools

AWS

GCP

Azure

Job description

We're hiring a Forward Deployed Specialist to sit inside enterprise accounts and turn AI capability into software that actually runs. You'll spend your weeks split between the customer's problem and the codebase: scoping messy workflows with the people who live in them, building the integration, then staying on it until it holds up in production.

Our client is an applied AI company working with large enterprises in complex, regulated industries where the data is fragmented, the security review is real, and nobody gets to hand-wave the last 20 percent. Forward deployed is not a support function here. It is the delivery model, and the people doing it carry both the technical judgment and the client relationship.

If you've been the engineer who gets pulled into the room because a demo impressed someone and now it has to be real, this is that job with the title to match.

What You'll Do

• Embed with enterprise customers to map their workflows, data, and constraints, then design the thing that actually fits
• Build and ship LLM-powered applications and agentic workflows across retrieval, tool integration, orchestration, and evaluation
• Wire into customer systems that were never built for this, including legacy APIs, on-prem data stores, SSO, and enterprise auth
• Run the technical side of pilots and expansions, from the first workshop through production cutover
• Write evals that tell you the truth about whether a system is good enough to hand over
• Translate between the customer's business language and the engineering team's, and be trusted by both
• Feed real deployment friction back to product so the platform stops requiring heroics
• Handle security reviews, data handling requirements, and procurement questions without stalling the build

Qualifications

• 5+ years building and shipping production software, with strong Python and solid API and systems fundamentals
• Comfort deploying and operating on AWS, GCP, or Azure, including containers, observability, and debugging code you did not write
• Integration experience across enterprise systems, data platforms, and identity and auth

Client-Facing Delivery

• Track record in a customer-facing technical role: solutions engineering, consulting, professional services, or forward deployed work at a product company
• Able to run a room with mixed technical and business stakeholders, executives included, and leave with a decision
• Judgment about scope, meaning you know what to build, what to defer, and when the honest answer is that AI is the wrong tool

Nice to Have

• Experience in regulated environments such as financial services, healthcare, energy, or public sector
• Background with MCP, agent frameworks, or internal tooling that other engineers actually adopted
• Time at a startup where the delivery playbook did not exist yet and you wrote it

Why Join Us

The team is small enough that you will shape how deployments get done, not just execute someone else's runbook. Engineers here sit close to the customer by design, and the bar is set by whether the system is still running and still used three months after go-live, not by whether the pilot demoed well.

You will get real ownership of accounts, direct access to the people building the platform, and the kind of pattern recognition that only comes from putting AI in front of users who are not impressed by it yet.