Build Antibrittle Agents With Us

https://jobs.gem.com/southbridge-ai

Any expression of intelligence is a transformation of data.

At Southbridge, we make that transformation reliable. We are a team of engineers and researchers building the first agentic data layer—systems where long-horizon AI agents can ingest, structure, and retrieve enterprise data without shattering. We call these antibrittle agents: orchestration stacks that get stronger the longer they run.

Why Context Comes First

AI models keep getting sharper, but most still feel like brilliant interns who can’t find anything. The bottleneck is context. Vector stores slice documents into fragments. ETL scripts drift out of date faster than they’re written. Enterprises teach new team members the shape of their data over six-year journeys—then watch the schema change.

We build for the people who live that frustration. Our walking-RAG pipelines read 1,000-page manuals, engineering diagrams, and telemetry streams without losing their relationships. Tadpole, our long-horizon runner, keeps Claude-class models productive for hours by trenching state, resetting cleanly, and capturing receipts with every step.

How We Think

• Context, not chunks – Every table remembers its document; every value knows its column. We keep structure intact from ingestion to retrieval.
• Agents, not brittle pipelines – Static workflows crack when reality shifts. Our agents reason about ingestion, transformation, and retrieval in real time and adapt when schemas move.
• Receipts above claims – We care about “nines of accountability.” Costs, decisions, failures, recoveries—everything is inspectable, everyone holds the same bar.
• Building blocks, not black boxes – We publish what we can under Apache 2.0 because shared primitives help us move faster and keep us honest. Offmute, diagen, wishful-search, tip20, next-cursor-base - they all started as tools we needed first.
• 180s are expected – Models evolve weekly. When the data asks us to change direction, we do it together.

What You’ll Work On

• Agentic orchestration – Extend Tadpole’s execution architecture so agents coordinate ingestion, schema inference, and retrieval with accountable state. You’ll do this alongside the people who sketched the first trenches.
• Context-preserving retrieval – Move past chunks. Build walking pipelines that understand nested JSON, spreadsheets, and industrial PDFs end-to-end, with teammates who have scars from legacy ETL.
• Human-first interfaces – Surface agent decisions with clarity: live traces, streaming diffs, receipts that help partners trust what they see, and that help us debug together.
• Reusable primitives – Extract the tools we need into open-source micropackages—utilities that the wider AI community wants to fork tomorrow, and that we’re proud to support.

Your first few months are collaborative by design. You’ll co-own a meaningful slice of Tadpole (backend) or southbridge.ai (frontend) with another teammate, ship it to real users, and share at least one internal note or open-source artifact. We’ll bring you into customer deployments early so you see the stakes up close, with the rest of us on the call.

What We’ve Already Built

• Antibrittle agent runners that keep Claude, Gemini, and Kimi K2 productive for multi-hour tasks.
• Walking RAG systems that read technical manuals, diagrams, and mixed-format archives for Fortune 150 teams.
• Open-source tooling with real adoption—Lumentis, offmute, diagen, tip20, wishful-search, and more.
• Evaluation harnesses that track new model releases the week they land, so we can pivot with evidence.

Front End Engineer

Full-time · Founding Team

https://jobs.gem.com/southbridge-ai

We’re turning live agentic data operations into experiences people can see, trust, and steer. As our frontend engineer, you will define how Tadpole’s long-horizon agents present themselves—on southbridge.ai, inside customer consoles, and throughout the documentation we publish. You’ll sit alongside the engineers building the orchestration core so the interface and the engine evolve together.

What You’ll Shape

• Product interfaces for autonomous AI – Build real-time views into ingestion, schema inference, and retrieval runs. Users should understand agent intent, confidence, and receipts at a glance, with backend partners feeding you the signal you need.
• Living data visualizations – Show nested structures, relationships, and deltas across thousand-page corpora. Help people spot what changed during a four-hour run, with data engineers ready to explain every curve.
• The face of Southbridge – Own southbridge.ai, interactive explainers, and launch materials that make “agentic data layers” intuitive to engineers, execs, and partners. We’ll support you with stories straight from live deployments.
• Docs & trails – Create Lumentis-style documentation, walkthroughs, and examples that sit alongside working code. Ship artifacts developers can clone and extend, with the rest of us reviewing for clarity and correctness.
• Internal tooling – Craft evaluation dashboards, diff viewers, and debugging consoles the team relies on daily. Expect to iterate with the people using them most.

Across your first few months, you’ll partner with another engineer to refresh southbridge.ai, deliver a streaming run console that ships with Tadpole, and publish at least one interactive doc or demo that earns community reuse. We’ll shape the roadmap together and make sure you have the context you need.

The Canvas & Tooling

• React + Next.js (RSC, streaming, edge rendering) with full TypeScript coverage.
• Tailwind, custom design tokens, and animation libraries where motion clarifies agent behavior.
• D3.js, Recharts, or bespoke canvas work for large-scale data visualizations.
• Real-time transport over WebSockets, Server-Sent Events, and experimental streams straight from Tadpole.
• Build pipelines shared with the backend team: Vercel deployments, Playwright smoke passes, visual regression where it counts.
• Everyday tools: Claude Code, Cline, Gemini CLI, Cursor, v0, our component workbench, and an internal design crit each week.

How We Work With You

• Designers are you—we expect taste, spacing discipline, and the ability to narrate decisions, and we make time for design reviews so you’re never alone in that call.
• You sit next to backend and infra peers (in person or timezone-close) to shorten the loop between capability and UI, with regular pairing time on the calendar.
• We prototype in public. Nothing ships without interactive receipts or at least a story of why this interface helps users trust the agent, and we’ll help gather those receipts.
• “Wide is better than deep” applies here too: we iterate on design systems as models change, not twice a year, and we compare notes as we go.

You Might Thrive Here If

• You’ve shipped complex TypeScript frontends where correctness and performance both mattered, ideally with a tight feedback loop to the people using them.
• You can read a JSON schema or API trace and immediately imagine how to visualize it, then talk through trade-offs with whoever produced the data.
• You’re comfortable owning the stack from lint rules to deployment, and you set guardrails before glitter, inviting the rest of us to test along the way.
• You’ve built interfaces for AI products—streaming chats, diff viewers, prompt builders, evaluation consoles—and you have thoughts on what made them humane.
• Words matter to you. Microcopy, onboarding flows, and empty states are product work, not afterthoughts, and you enjoy pairing with product and research on them.

Signals We Love

• Open-source components, design systems, or tools others already rely on—and notes about what you’d improve in version two.
• Screenshots or recordings of dashboards you’ve crafted that balance detail with calm, plus a sentence about the constraints you managed.
• Stories where you persuaded teams to delete complexity and explain things visually instead—and what happened after launch.
• Curiosity about the models themselves—you run evals, test new releases, and adjust UI affordances when behavior shifts, often in conversation with the rest of the team.

Why This Work Matters

You’ll define new interface patterns for long-horizon AI. When enterprises trust their data to our system, it’s because your interface helped them understand and believe the agents they’re watching. Technical depth meets craft here—we know great UX multiplies the value of the infrastructure we’re building, and we ship to real users fast so your pixels turn into decisions within days.

Hiring Process

1. Please share your Portfolio by sharing a short Loom/Tella/Video platform of your choice walk-through. We work globally so remote async communication skills are important.
2. Include links, repos, production screenshots, or Show us what "taste with teeth" means to you and tell us the context behind it.
3. Meet one of the team – 30-minute conversation to see if the work resonates and answer your initial questions.
4. Conversation with Hrishi – 45 minutes on narrative, craft, and how you think about AI UX; bring questions about how we work.
5. Take-home challenge – We'll hand you a Tadpole trace or interface problem to work through. Realistic constraints, time to think.
6. Interface deep dive – Walk through your solution end to end, explore trade-offs together. We'll also review a prior project of yours to understand your approach.
7. Your questions for us – Dedicated time for you to interview us on anything that matters to your decision.
8. Pairing session – Co-create or refine a live component in our codebase with the team. Expect Tailwind, streaming data, and honest feedback both ways—we'll share the brief beforehand.
9. Offer & runway – We tune scope, compensation, and relocation support so you can operate at full pace, and we map out the support you'll have on day one.

Ready to give long-horizon agents a face people can trust?

Apply below with something you’ve shipped that still makes you proud—and, if you like, a quick sketch of how you’d show an agent walking through a document.

https://jobs.gem.com/southbridge-ai