Summer 2026 - Internship

Build the AI Workforce for Financial Operations

At PathPilot (YC S24), we believe the future of software isn’t dashboards — it’s agents that do the work.

We’re building a system of AI agents that interact with customers, make decisions, and collaborate with each other to execute real workflows across fintech companies — from collections and support to onboarding and compliance.

Not copilots.
Autonomous systems that run operations.

---

Why now

Even today, a large share of lending operations — often ~70–80% across the US and LATAM — is still manual:

People clicking through CRMs, following scripts, chasing payments, verifying documents.

Not because it’s simple — but because it’s:

• High-stakes (money + risk)
• Fragmented across systems
• Constrained by compliance

That’s exactly what makes it the perfect problem for AI agents.

---

What we’re building

A network of agents that:

• Talk to customers across chat, voice, WhatsApp, email
• Negotiate payments and track commitments
• Verify identities and documents
• Enforce compliance in real time
• Coordinate with each other to complete workflows end-to-end

These systems are already driving real outcomes (cost reduction, higher recovery, faster ops) in production environments.

---

Why PathPilot

We’re not building demos — we’re deploying agents inside real fintech operations.

• YC-backed (S24)
• Customers handling tens of thousands of interactions/month
• Focused on one of the hardest verticals: lending operations

Founders

• Built and scaled production systems before (ex-Facebook engineering, repeat YC founder)
• Saw firsthand how broken and manual financial operations are
• Now building the AI layer to replace them

---

Join us (In-person)

As a Fullstack Intern, you’ll work directly with the founders to build this system.

Stack: Python + TypeScript

---

What you’ll actually do

This is not a “tasks from Jira” internship.

You’ll own real parts of the system, for example:

• Building the TypeScript integration layer connecting agents to CRMs, dialers, and financial APIs
• Designing agent workflows (multi-step, multi-agent coordination)
• Improving reliability (evaluation, retries, guardrails)
• Shipping features used in live customer environments

---

First 30 days

By the end of your first month, you will have:

• Shipped an end-to-end agent workflow
• Example: an agent that handles automated payment plan negotiations or resolves support tickets
• Deployed it into a real environment (or close to production)

---

Who this is for

You should apply if:

• You’re a strong builder (TypeScript + backend)
• You’ve built something with LLMs (agents, RAG, etc.) — not just tutorials
• You like owning problems end-to-end
• You want to move fast, in-person, with high responsibility

---

How we evaluate candidates

We prioritize people who have:

• Built and deployed real projects (especially AI/LLM-based)
• Shipped things end-to-end (not just coursework)
• Strong signal of taste + ownership + speed

If you’ve built something scrappy that works > perfect academic projects.