About Humanize

Behind every business decision is a person - what they want, what they would tell a friend, or why they quietly left. But actually talking to them has always been too slow and expensive to do often. Everything else is a guess.

AI started to change that - but most tools stopped at surveys or one-off interviews. Useful, but lacking interaction. Real insight comes from conversation: people riffing off each other, challenging assumptions, saying the thing they didn't plan to say.

Humanize powers AI-moderated group conversations with real people and translates what they said into the tools where decisions actually happen. We're building a world where the customer's voice is always in the room - not only when someone commissions a study.

Where We Are

Funded. We just closed our pre-seed, with participation from Sovereign's Capital, Hearst Ventures, and Cubit Capital, plus angels including Jason Karp, Brett Hurt, Dan Kloiber, and Jesse Pujji.
Founders. Multiple prior exits between us, including companies built to $1B+ valuations.
Traction. Live with paying customers and growing rapidly month-over-month. We're heads-down on product, not press - happy to share specifics under NDA.

Location & Logistics

Hybrid / Remote-friendly - NYC office overlooking the city near Columbus Circle, with a distributed team
Reports to: CTO
Visa: We're not able to sponsor work authorization at this time

About the Role

We're looking for a Senior Software Engineer who ships with urgency but doesn't cut corners. Someone who spots a bug, opens a ticket, and fixes it - without waiting for anyone to ask. Someone who resolves ambiguity before it slows the team down. You'll own features end-to-end, from system design to production.

What You'll Do

Build and improve the AI voice agent that moderates live group conversations - turn-taking, cross-talk, silence, and the moment someone says the thing they weren't going to say
Work on real-time systems - live audio/video, per-participant transcription, session orchestration
Extend our AI analysis pipeline - sentiment, theme extraction, and structured output the rest of the company actually uses
Build the agents that find the right people to talk to - recruitment, screening, and panel quality
Evolve our data layer - migrations, security policies, type-safe query patterns
Own third-party integrations
Use AI coding agents as a core part of your workflow - we treat these as multipliers, not crutches
Collaborate directly with product and design to scope and ship work
Review code, surface risks early, and help the team move faster

Technical Challenges

AI moderating a group, not a 1:1. Most "AI interviewers" handle one person at a time. We're orchestrating five - managing turn-taking, follow-ups, drift, and the unprompted moment that's the whole point of qualitative research. Real-time, multi-speaker, low-latency.
Recruitment as an agent. Given a research goal, find the right participants - not just on paper, but ones who'll actually show up and engage. Sourcing, screening, no-show prediction, panel health over time.
Sentiment, not surveys. Reading video, voice, and language together to surface what people meant, not just what they said - and routing it into the systems where decisions actually get made.
Conversation → decision. Turning hours of group conversation into structured outputs that flow into Linear, Notion, Slack, and the dashboards the customer already lives in.

What We're Looking For

Technical

5+ years building production software
Strong TypeScript and React - you've built complex, stateful UIs beyond CRUD forms
Experience with Python (used in our realtime AI agents)
Solid PostgreSQL skills - migrations, indexing, row-level security, JSONB
Experience building and maintaining APIs (REST or RPC)
Comfort with real-time systems (WebSockets, WebRTC)
Familiarity with LLM/AI integration in production - prompt design, streaming, structured extraction

As a Person

Negative maintenance. You make life easier for everyone around you. You fix things before they become problems. You ask yourself: what can I do today that will reduce future work for others?
Non-attached action. You ship the functional version first and iterate - no gold-plating v1. You'll build something knowing it may be rewritten, and when the evidence says pivot, you pivot without ego.
Proactively curious. When something doesn't make sense, you dig in and challenge it - not shrug and move on.
Reliable. If you said you'll do it, it's getting done. No news is good news - and if it's not, you raise the flag early.
Give a damn. You show up with energy, not just availability. You care about the product, the users, and the people you work with, and it's obvious.

Tech Stack

React / TanStack / shadcn / Hono / Kysely / Supabase / Vercel / LiveKit