Technical Co-Founder at Modern Collections

Company Overview Stage: Pre-seed → seed · Team: 1 (solo founder, hiring co-founder) · Stack: Python / FastAPI / Postgres / React · HQ: Remote (US)

What we're building The AI-native operating system for B2B commercial debt collections.

Commercial creditors — manufacturers, distributors, freight, SaaS vendors, professional services — are owed roughly $3–4B of placed receivables annually in the US, generating ~$300–400M in agency fees. They place that debt with collections agencies running on software from the 1990s: 90-day onboarding, faxed status reports, opaque compliance posture, and 70% of agency labor spent reading email and updating a CRM.

We replace the agency. Same contingency-fee model, fundamentally different unit economics.

What we run today One FastAPI monolith, five logical components, in production:

Compliance engine — federal Reg F + per-state rules engine (contact windows, license requirements, consent rules), with a deterministic veto over any outreach. Every decision is audited and expires after 24h. Voice agent — ElevenLabs + Twilio behind a single platform adapter. Barge-in, two-party-consent detection, dispute capture, full-call recording mapped to per-state retention rules. Email engine — Postmark-backed, compliance-gated, sequenced. Enrichment pipeline — debtor identity resolution, contact discovery, KYB + sanctions on the creditor side. Creditor portal + API — Clerk auth, Stripe payments, Postgres RLS for tenant isolation, append-only audit log. It's not a prototype. It runs end-to-end. We can place a debt and collect it without a human in the loop on the happy path.

Operating model: 3P first, 1P second 3P (Third-Party): we collect on behalf of creditors, take the compliance risk, take a contingency fee. This is the wedge. 1P (First-Party): creditors run our software themselves and take their own compliance risk. This is the SaaS arc, unlocked after we've earned the trust and outcomes data. We lead with 3P proudly. Bessemer and Khosla now prefer "service-as-software" over pure-SaaS for vertical AI because the services revenue funds the moat and the outcomes data trains the product.

Why now Regulation F (effective Nov 2021) federalized the rulebook for the first time in 40 years, ending the patchwork status quo that protected legacy agencies. Compliance is now codifiable. Voice AI crossed the line in 2024–2025. ElevenLabs + GPT-class reasoning + sub-300ms barge-in means a synthetic agent can hold a compliant collections call. Two years ago it could not. fintech infrastructure is finally horizontal. Stripe, Plaid, Middesk, sanctions APIs, and Clerk make a new entrant viable with a five-person team. The bill of materials is 10% of what it was in 2018. Legacy agencies are consolidating and shedding talent. TrueAccord, Symend, Prodigal restructurings have put domain-fluent engineers into the market. Why this is defensible Not "we use AI better." The moat compounds across five structural assets, each of which is a capital cost competitors must pay and a timeline they can't compress:

Per-state compliance config as code, version-controlled, audited per decision Multi-state licensing footprint — 21 licenses targeted by month 18 Append-only audit trail that survives an FDCPA lawsuit Voice agent + state-rules engine integration — barge-in, consent, recording retention all tied to the same config Outcomes data — every call, every payment, every dispute, in one schema, training the next quarter's prompts Where we are Solo founder (CEO), working platform in production already First creditor pilots underway Pre-seed raise in progress Actively hiring a technical co-founder — this is the most important decision of the next 24 months (and the reason you're reading this) What the next 24 months look like Months 0–6: First 10 paying creditors. Compliance engine hardened. State-rules engine generalized beyond the first few jurisdictions. Months 6–12: $5M ARR run-rate. Licensed in 8 states. Voice agent handles 50% of outbound volume. Months 12–18: $20M ARR run-rate. Licensed in 21 states. Begin 1P (self-serve) pilots with the largest creditors. Months 18–24: $20M → $50M ARR. Series A. 1P SaaS line generating real revenue. Who we are A solo founder today. The CEO owns commercial, GTM, fundraising, and has shipped the platform that exists in production right now. The open seat is for a technical co-founder — peer on the cap table, 25% equity, CTO title — who actually understands AR automation and commercial collections and will own the entire technical surface area from day one. We work remotely, in US time zones, and we ship every day.

The fact that this is a team of one is the opportunity, not the caveat: you walk in as co-founder #2 with the platform already shipping, a fundraise in motion, and uncontested ownership of the entire technical org you'll go on to build.