About Aboon
Setting up a 401(k) shouldn't be the hardest thing a business ever does. But retirement plan administration is complex, fragmented, and manual — and it slows down both business owners and the advisors who serve them. Aboon was founded to change that.
We're an AI-native third-party administrator (TPA) that makes it easy for advisors to offer 401(k)s and other retirement products by pairing streamlined workflows with dedicated human service. Backed by leading investors including Bain Capital and Edward Jones Ventures, we bring decades of experience building for advisors and the retirement industry. We're looking for exceptional, mission-driven people to build it with us.
About the role
We're seeking an applied AI engineer to help us automate 401(k) administration — building not just tools for administrators, but the agents and workflows that do the work, with our experts reviewing and approving the results.
You’ll be working in a pod, paired up with a retirement plan expert (the SME) and collaborating in a daily loop: reviewing flagged sessions, adjusting agent procedures, tuning thresholds, improving our evals, and deploying agent updates to production continuously. We aim to measure feedback in hours, not days, and are building a harness tailored explicitly to our servicing team — the tools, procedures, eval gates, and audit trails that let agents act safely inside a complex fiduciary business.
Much of what you ship will be deterministic workflows with AI at the edges: reliability first, autonomy earned through evidence, and agentic loops reserved for work whose steps can't be enumerated in advance. You’ll help the team make structural changes to our harness through our Shape Up-inspired cycles and have an eye towards the future as we continue to scale our agentic operation.
You'll join a small, tight-knit team in a flat hierarchy and report to our Head of Engineering.
What you’ll do
Build the harness. We’re building a harness that lets us rapidly deploy and tune agentic workflows and automations. You’ll help identify the tools we need, pitch structural improvements, and build an accessible evaluation loop for our SMEs.
Convert our experts’ SOPs into agent operating procedures (AOPs). Versioned, with explicit findings and per-step automation levels; not just scattered prompts.
Run the daily loop with your pod's expert. Baseline all new agents against golden cases, hillclimb the judgement steps, expand the evaluation data set, and deploy agents when they’ve met the bar.
Own our audit and observability stack. Agent input, output, and human decision making logged and compliance-ready.
Build the testing layers that keep agents trustworthy. Replay datasets against every change to catch regressions. Monitor the external sites and services we automate so we know when they change under us.
Expand automation gradually. Once humans have reviewed a step's output repeatedly without changing it, that step stops requiring review. You build the tracking that makes those calls data-driven.
Our stack
Python agents on Agno; Langfuse for traces, evals, and annotation.
Reducto for document processing; Skyvern for browser automation.
Rails/PostgreSQL as our primary platform.
Redshift + dbt + Metabase; PostHog for product analytics.
HubSpot, Asana, Slack, and GDrive as the surfaces where agent work meets human review.
Who you are
Strong communicator: You’ll be sitting alongside non-technical domain experts and helping encode their judgement into our workflows and agents.
Pragmatist about AI: You have a sense for when to reach for an agent loop, a deterministic step, or a human gate.
Operator at heart: You keep close tabs on metrics and sessions, and you are proactive in addressing issues as they arise.
What you bring
Required
5+ years of software engineering experience, with 2+ years building LLM-powered systems in production.
Strong Python; comfortable working alongside a Rails monolith on AWS.
Experience with an agent runtime (Agno, LangGraph, or similar) and eval tooling (Langfuse or similar).
Proficiency working with coding agents.
Experience building golden datasets and gated releases on eval results — or you're ready to make that your default motion.
Nice to have
Experience in regulated or adjacent industries (fintech, insurance, healthcare). Retirement/401(k) domain exposure extra nice to have.
Have built browser automation at production scale, and/or document extraction over messy real-world inputs.
What we offer
Compensation
The base salary range for this role is $220,000 - $240,000, depending on experience, skills, and location. Total compensation also includes an annual bonus and equity.
Benefits
Comprehensive health, dental, and vision insurance, with additional optional coverage
401(k) with company match
Flexible paid time off and holidays
10 weeks paid parental leave
A fully remote work environment, with all the technology and equipment you need
Support for professional development and career growth
Location
Remote, though we encourage in-person attendance at 1-2 company retreats per year. This role is open to candidates based in the US. Work authorization required; no visa sponsorship.
Aboon is an Equal Opportunity Employer. We encourage applications from all qualified individuals and do not discriminate based on race, color, religion, sex, sexual orientation, gender identity, national origin, disability, veteran status, or any other protected status.