Software Engineer - AI Copilot

Assembled builds the infrastructure that underpins exceptional customer support, empowering companies like CashApp, Etsy, and Robinhood to deliver faster, better service at scale. With solutions for workforce management, BPO collaboration, and AI-powered issue resolution, Assembled simplifies the complexities of modern support operations by uniting in-house, outsourced, and AI-powered agents in a single operating system. Backed by $70M in funding from NEA, Emergence Capital, and Stripe, and driven by a team of experts passionate about problem-solving, we’re at the forefront of support operations technology.

Our Copilot team is building the bridge between human support agent and AI in customer support. Think “Cursor for Copilot”: an industry-leading, assistive experience that helps agents write, reason, retrieve knowledge, and automate the rote—so they can focus on judgment and empathy. This is a strategic investment for Assembled: our vision is to elevate support with AI, not replace humans. We’re pushing beyond AI-only competitors by making human-in-the-loop workflows fast, trustworthy, and a joy to use.

• Wide, diverse users: Build for many agent personas—different case types, tones, writing styles, tech comfort, and AI fluency—like a mash-up of consumer polish with B2B power.

• Raise the bar for “assisted,” not “sloppy autonomous”: We’re crafting crisp human-in-the-loop experiences where AI augments judgment rather than undermining it.

• Real impact at scale: Our AI copilots are already used and loved by thousands of agents daily (and the number is increasing rapidly), improving quality of life by automating the repetitive and spotlighting critical thinking and empathy.

• Cutting-edge GenAI: Apply LLMs for retrieval, complex reasoning, and action automation—grounded in real production constraints of latency, reliability, and safety.

As an AI Copilot Software Engineer, you’ll ship product that agents love using all day: composing responses, summarizing threads, retrieving facts with context, suggesting next actions, and safely executing automations. You’ll prototype quickly, run experiments, and harden what works into reliable, scalable features.

• Build end-to-end Copilot features: drafting, summarization, translations, tone/brand alignment, and one-click workflows.

• Implement retrieval & reasoning pipelines (RAG, tools/actions), balancing quality, latency, cost, and safety.

• Instrument, evaluate, and iterate: define success metrics, run A/Bs, and evolve prompts/models with telemetry and evals.

• Partner with users and customers: shadow agents, synthesize insights, and turn feedback into shippable improvements.

• Collaborate across product, design, and infra to take ideas from sketch → prototype → production → monitoring.

• Contribute to our Copilot platform: shared evaluators, prompt libraries, offline/online testing, and guardrails.

• Have 3+ years as an individual-contributor engineer (startup/small-team impact a plus).

• Are fast-moving and love rapid iteration/experimentation.

• Are product-minded and customer-obsessed; you measure success by agent happiness and business outcomes.

• Have strong fundamentals in a modern language (Go, TypeScript, Python, etc.) and comfort across backend or full-stack.

• Thrive in ambiguity, own outcomes, and communicate crisply.

• No prior AI experience required (nice to have)—you’re excited to learn and ship.

• Experience with LLMs (prompting, RAG, tool use), evals/observability, or safety/guardrails.

• Knowledge of support platforms (Zendesk, Salesforce, Freshdesk) or enterprise integrations.

• An taste for UX polish and accessible, trust-building UI for assistive features.