Mid-Level Software Engineer, Chicago (remote or in-person)
Impruve is a boutique forward-deployed engineering firm that helps wealth management firms get real business results from AI. We embed directly with our clients to build production data pipelines, AI tools, custom software, and compliance infrastructure for firms managing billions in client assets. We have more demand than we can currently serve, and we're growing the team to meet it.
Our founders previously built and exited Packback, an AI company that served more than 3 million users.
WHO WE ARE
We're four engineers right now, and we're building how we work on purpose, not backfilling a culture after the fact. We care about doing right by clients and by each other. We take the work seriously: real client money, real compliance risk, real production systems. We don't take ourselves too seriously.
We pair early and often, across the team, because it's how we actually get better. If you notice something interesting mid-task, we want you to flag it and pull someone in, not disappear down a rabbit hole solo. When something breaks, we run blameless postmortems; the point is fixing the system, not finding someone to blame.
We don't expect you to walk in with all the answers. Nobody here has them either. We expect you to be scrappy: willing to try something, see what happens, and figure it out with the team rather than wait until you're sure.
You won't be joining something fully formed. You'd be joining our founding engineering team, with real say in how this team works, not just what it ships. Ask us about that in the interview and we'll actually tell you, not recite a values poster.
WHY THIS ROLE
Most wealth management firms want AI in their business and have access to the frontier models. What's missing is the engineering capability to deploy it effectively and compliantly. That's the gap Impruve fills.
This is the role where you learn that craft by doing it. You'll own real workstreams inside live client engagements, with a senior engineer owning the engagement itself, so you have real air cover while you build judgment about shipping AI into production environments that matter.
WHAT YOU'LL DO
- Own a real workstream inside a live client engagement, spec to production: the senior engineer holds the overall relationship, the outcome of your piece is genuinely yours.
- Pair regularly with teammates across the stack, not just when you're stuck. It's the normal way we work, not a last resort.
- Move across the stack: data pipelines, RAG and agent workflows, evals, backend, frontend, and learn the AI-specific pieces on the job if you haven't done them yet.
- Ship AI that survives contact with real users and real compliance reviewers, not demos.
- Work AI-natively: spec-driven development, custom agent skills, Claude Code, and the rest of the AI tooling ecosystem are just how we build. Bring what you learn back to the team.
- Help set the norms we're all still figuring out, from how we review code to how we run postmortems.
WHO THRIVES HERE
You would be a great fit if you identify with these:
- 2 to 5 years of hands-on engineering experience shipping and operating production systems real users depended on.
- Real depth in two or more of backend, frontend, data engineering, infrastructure, or AI/ML.
- Comfortable not knowing the answer yet: you'd rather try something and see what breaks than wait until you're certain. You don't need prior experience with agentic workflows or RAG pipelines specifically; scrappy and curious beats a resume line here.
- Collaboration energizes you rather than draining you. You'd rather pair through a hard problem than sit on it alone.
- You want a hand in shaping team norms, not just following ones already set.
- You communicate clearly with engineers and non-engineers alike, including clients.
- You're conscientious about the unglamorous stuff: testing, monitoring, observability.
- You use AI as a tool to work more efficiently; you do not use it to replace original thinking, creativity, and human collaboration.
Stack: Python, Node.js, Express, TypeScript, Next.js, MongoDB, Redis, AWS ECS, Docker, LangChain/LangGraph. We care more about sound judgment across domains than a checklist match.
HOW WE HIRE
Three steps: an initial conversation, a team interview with the founders that includes a technical problem-solving session, and a final decision.
ONE MORE THING
Research consistently shows that women and people from underrepresented groups are less likely to apply to a role unless they check every box, while others apply anyway. If that's you: apply anyway. We care more about curiosity, scrappiness, and kindness than a perfect resume match, and we build a better team when the people on it don't all think alike.
Requires US work authorization; we can't sponsor visas.