Careers
Help build the underwriting OS for small business lending.
We're a small, focused team building agentic AI for one of financial services' most operationally painful workflows: small business lending. If you want to work on real customer problems, messy data, documents, AI agents, credit workflows, and software that ships into production, we'd love to talk.
Small business lending still runs on PDFs, inboxes, and human heroics.
Every day, lenders receive messy borrower packages: applications, bank statements, tax returns, invoices, IDs, SOS records, emails, and CRM notes. Before a credit decision can happen, someone has to organize the file, verify the business, classify transactions, detect fraud signals, apply policy, and write the memo.
Kaaj exists to turn that work into structured, auditable credit intelligence — so lenders can move faster, serve more businesses, and let humans focus on judgment instead of manual assembly.
Real-world AI
We work on documents, reasoning, extraction, verification, memory, and agentic workflows that have to survive messy production data.
High-stakes fintech
Our users make credit decisions. Accuracy, auditability, traceability, and trust matter.
Operational leverage
Every workflow we automate gives lenders more capacity without simply adding headcount.
Why join now
Early enough to shape the company
You will not be maintaining a giant machine. You will help define the product, engineering culture, customer experience, and internal standards from the ground up.
Late enough to work on real demand
We are not searching for a problem. Lenders are already asking for faster, more reliable ways to process small business credit.
A rare applied-AI surface area
This is not chat on top of a database. The product has to understand documents, business context, credit policy, anomalies, exceptions, and human review loops.
Close to customers
You will see how banks, lenders, brokers, and credit teams actually work — and ship improvements that change their daily operating rhythm.
Problems worth solving
Here are some of the concrete problems you could work on.
Document intelligence
Classify, extract, rename, and cross-check borrower documents across applications, statements, tax returns, invoices, IDs, and financials.
Agentic workflows
Design agents that observe a deal, branch based on context, gather evidence, and preserve their reasoning.
Bank statement analysis
Separate true revenue from transfers, refunds, MCA proceeds, owner injections, and one-time deposits.
Business verification
Combine Secretary of State records, web presence, domain age, address signals, ownership clues, and industry context.
Credit memo generation
Turn analysis into clear, editable, source-backed memos that credit teams can trust.
Enterprise integrations
Push structured outputs into Salesforce, LOS/CRM systems, file stores, and lender workflows without forcing rip-and-replace.
Security and reliability
Build infrastructure that handles sensitive borrower data with the discipline financial institutions expect.
Who thrives at Kaaj
You like ambiguity, but not chaos.
You can turn vague customer pain into clear product decisions and shippable software.
You care about craft.
You notice when an interface feels confusing, an extraction is brittle, or a workflow creates unnecessary manual review.
You move fast and own outcomes.
We value people who can go from problem to prototype to production without waiting for perfect instructions.
You are customer-curious.
You want to understand how underwriters, brokers, analysts, and operations teams actually work.
You are rigorous with AI.
You do not treat model output as magic. You think about evals, edge cases, fallbacks, confidence, and human review.
You communicate clearly.
Written clarity matters here because we build products that turn messy information into decisions people can trust.
How we work
Start from the file, not the demo.
We care about what happens inside real borrower packages, not just polished product flows.
Automate the science. Preserve the art.
The goal is not to remove human judgment. It is to remove the manual work that prevents good judgment from scaling.
Reasoning should be visible.
In credit workflows, answers are not enough. Users need to know why the system concluded what it did.
Ship with evidence.
We like fast iteration, but production systems need tests, evals, observability, and customer feedback.
Earn trust in the details.
Security, traceability, accuracy, and workflow fit matter when your users are lenders.
Open roles
We are always interested in exceptional engineers, designers, and GTM operators. Tell us what you want to build.
Founding Full Stack Engineer
Build core product workflows across document intake, agent orchestration, customer-facing UI, integrations, and internal tools.
Founding Full Stack Engineer
Build core product workflows across document intake, agent orchestration, customer-facing UI, integrations, and internal tools.
Founding GTM
Work with lenders and brokers to understand workflows, support implementation, refine messaging, and turn customer learning into growth.
Use the form above to submit your LinkedIn/GitHub/portfolio, a short note on why Kaaj, and one project you are proud of.
What to expect
Intro conversation
A short conversation about your background, interests, and what Kaaj is building.
Work sample or technical discussion
Depending on the role, we will discuss a relevant project, product problem, technical design, or practical work sample.
Founder / team conversation
You will meet the people you would work with and get a clear view of how we operate.
Reference and offer
If there is mutual fit, we move quickly.
Where we work
We are based in San Francisco and work with teammates who can collaborate across US-friendly time zones. For early roles, we value people who can spend meaningful time close to customers, product decisions, and fast feedback loops.
San Francisco
Our home base. Ideal for roles that benefit from frequent founder and customer collaboration.
Remote
Remote-friendly for high-ownership candidates who communicate clearly and overlap with US time zones.
Customer-facing
Some roles may involve travel to customers, conferences, or onsite implementations.
Benefits and support
We are early, so our benefits will evolve as we grow. We aim to be thoughtful, transparent, and competitive for the stage of the company.
Customer impact
“Kaaj's AI agents have become a force multiplier for our operations, streamlining intake and credit workflows, eliminating manual data entry, and empowering our team to respond to partners faster and more consistently. The result: a sharper, more reliable experience at every touchpoint”
Want to build with us?
If you are excited about AI, fintech, and real operational problems, reach out. We care more about ownership, clarity, and craft than perfectly matching a job description.
Apply — Engineering