$12 trillion in goods move by freight every year, coordinated by people copy-pasting between email and spreadsheets. We’re building the AI that changes that.

Sidecar builds AI to run the global supply chain.

Ever wonder how your phone got to you from a factory in Shenzhen? Or how the jacket you ordered from Italy made it to your door in 5 days?

Behind every shipment is a freight forwarding company, a logistics middleman coordinating ships, customs, warehouses, and trucks across continents. These companies still run on emails, spreadsheets, PDFs, and terminal software from 1998. Their teams manually extract info from documents, and respond to hundreds of emails and phone calls daily to execute business specific workflows.

Sidecar is a 24×7 Virtual AI Teammate for these Logistics Service Providers that calls, emails and logs into systems and everything else to execute a complex task.

We are backed by leading Silicon Valley VC’s, Angel Investors and Industry Advisors and pay top-bracket salaries alongside equity compensations.

The Role

You will own the surfaces that humans actually touch. The operator dashboards, the email and chat plugins, the internal tools, and the APIs that glue them to our AI agents. Our agents process 1M+ webhooks per week across 20+ freight forwarding companies, and you build the product layer that turns that raw intelligence into something a logistics operator can use in seconds.

This is not a “ship a button” role. You will write TypeScript and Python, design schemas, build APIs, debug production issues across the stack, and own features end-to-end, from spec to a customer using it the same week.

Here’s what you’ll work on

• Client-facing web app: dashboards, queues, review and HITL surfaces, analytics. Built in Next.js, React, TypeScript, Tailwind.

• Plugins and integrations: Outlook add-in, Microsoft Teams chatbot, in-product chat surfaces. Wherever the operator already works.

• Backend APIs: design and ship REST endpoints in Python/Flask, model Postgres schemas, write migrations, optimize hot paths.

• AI pipeline integration: wire LLM outputs (draft generation, classification, extraction) into operator surfaces. Handle latency, partial responses, fallbacks.

• Customer-grade UX: build interfaces a freight operator can learn on sight, with no docs.

• Testing and reliability: vitest, React Testing Library, Playwright, eval gates on prompts powering the UI. The bar is “no regressions reach production”.

• Production debugging across the stack: a state bug on the frontend, a Postgres slow query, an LLM hallucination, a payload mismatch. You find which layer is lying and fix it.

Qualifications

• Frontend (3+ years production). Next.js or React, TypeScript at scale, modern state management, Tailwind or equivalent. You’ve shipped production web apps used by real customers, not just internal dashboards your team uses.

• Backend (2+ years production). Python (Flask/FastAPI/Django) or Node. You’ve designed REST APIs, modeled schemas, written migrations, and debugged production incidents. Not just “I called an API from the frontend”.

• PostgreSQL. Schema design, indexing, query optimization. You’ve fixed a slow query under load and know what EXPLAIN ANALYZE says.

• End-to-end ownership. You’ve shipped features from spec to production: frontend, backend, DB, deploy. Without a PM, a designer, or a QA team holding your hand.

• Testing discipline. vitest/Jest, RTL, Playwright or equivalent. You write tests because the alternative is a 2 a.m. page.

• Production debugging. You’ve SSH’d into a box, read a Sentry trace, and shipped a same-day fix.

Nice to have

• LLM API experience (OpenAI, Anthropic, Gemini): prompt design, streaming, latency budgets, fallback handling

• LangGraph or similar AI orchestration frameworks

• Plugin and integration development: Microsoft Graph API, Outlook add-ins, Teams bots, Chrome extensions

• Real-time UX: WebSocket, SSE, optimistic updates

• shadcn/ui, Radix, design-system work

• Docker, GCP or Azure (Cloud SQL, VMs, Container Registry)

• Celery + Redis task queues

• Sentry, Grafana, Prometheus

This is NOT the role if

• You only want to do frontend, or only backend. The role is both, in the same week.

• You need a designer to hand you a mockup before you can start. Half the time you’ll be the one deciding what the UI should look like.

• You think “the operator should figure it out”. Our UX bar is no training, no docs.

• You’ve never shipped a product to real, paying users.

• You prefer rewrites over fixing the existing thing under load.

Experience

3–5 years of relevant experience in full-stack product engineering at a startup or product company. Proven track record of owning customer-facing surfaces end-to-end. Experience with AI/LLM products is a plus but not required.

We work 996 out of our offices in San Francisco and Bangalore. We are highly selective and believe in only hiring and retaining the best.