Forward Deployed Engineer

Aviary•Chicago, IL

42d•Onsite

About The Position

A forward deployed engineer at AviaryAI takes a signed contract and turns it into a live AI agent making real calls and sending real texts to a credit union's members. Then you make it better until it hits the numbers the client signed up for. You sit between the client and the platform. You are the person who understands both what the credit union's VP of Operations is trying to accomplish and what the agent is actually doing on call 4,000. Most of the interesting problems live in the gap between those two things. This is a technical role with a client in the room. It is not a client success role with a technical veneer, and it is not a software engineering role where someone else handles the customer.

Requirements

Comfortable in the tools: JSON configs, SQL, reading application logs, understanding how data moves between systems through APIs and flat files.
The kind of person who opens the log file instead of filing a ticket.
Have shipped something real people used, and owned what happened after (a deployment, an integration, an internal tool, a data pipeline).
Lived through the gap between 'it works in testing' and 'it works for the customer.'
Can sit across from a credit union executive and hold the room. Explaining why the agent did what it did, what you are changing, and when they will see the difference.
Clear and specific, without hiding behind jargon.
Write clearly: Specs, runbooks, client updates, bug reports for engineering.
Half this job is making complexity legible to someone else.

Nice To Haves

One to three years in implementation, solutions engineering, technical support, or data and operations roles where you touched production systems
Python or another scripting language
Contact center, telephony, or conversational AI experience
Experience with core banking systems or financial data

Responsibilities

Build and launch agents.
Configure the agent for each client's use case: the prompt, the call flow, the transfer and escalation logic, the guardrails, the SMS fallback.
Test it against real scenarios before a single member hears it.
Own the integration surface. Work with the client's technical team to define the data spec, map fields from their core banking system, set up file delivery, and confirm that what comes back to them is what they need.
Read the logs and fix what you find.
Diagnose issues when an agent goes silent, interrupts a member, or takes too long to answer from the call logs and the data.
Run SQL against call health, isolate the failure, and either fix it or hand engineering a precise repro.
Own the outcome after launch. Align on success criteria before kickoff, then run optimization cycles on prompts, timing, and flow until the client's targets are met and hold.
Run the client relationship at the operator level. Weekly calls with the people who own the contact center and the outcomes.
Explain what the agent did and why, in language that makes sense to someone who does not think about LLMs for a living.
Make the next deployment faster. Turn what you learn into templates, checklists, and documentation.
Feed the product. What you bring back changes what gets built.