Software Engineer

Join us as a Software Engineer to build the AI execution layer where go-to-market agents do the work, not just answer questions.

Taizen builds purpose-built AI agents for Product Marketing, Sales Enablement, and Revenue teams. Our AI agents connect to tools like Gong, Salesforce, HubSpot, and Slack, then proactively run recurring workflows without waiting to be prompted., Go-to-market teams are drowning in signals they can't act on fast enough. Competitor moves, deal objections, buyer feedback, and coaching gaps are trapped across call recordings, CRM notes, documents, and Slack threads.

We're building agents that do the operational work:

• Competitive intelligence: Monitor competitor activity and calls, then deliver reports, updated battlecards, and rep talk tracks automatically
• Deal and account intelligence: Show sales leaders where deals are advancing or stalling, while helping BDRs research territories and identify the right accounts to bring into pipeline
• AI roleplay: Let reps practice against voice-based AI buyers grounded in real deals, objections, and battlecard content
• AI win/loss interviews: Interview buyers after a deal and turn the conversations into structured insights without someone scheduling every call, You won't just work inside this stack, you'll help decide where it goes. How our retrieval layer evolves as we grow past our first customers is still an open question you'd have a say in.
• Frontend: React
• Backend: two services, a Node/TypeScript core and a Python agent service
• Retrieval: our own vector store written in Rust, SOTA late-interaction retrieval
• Agents: our own harness, not an off-the-shelf framework, and model-agnostic
• ML: a mix of our own and open-source models; Modal for GPU work
• Infra: AWS (RDS, EC2, CloudFront, S3, ECR, ECS, ElastiCache, Lambda), MongoDB Atlas, Grafana + Prometheus, GitHub Actions

What you'll do

• Ship features end-to-end, from customer problem to design, deployment, and iteration - architecture decisions included
• Build reliable backend services and workflows behind the agents
• Build product surfaces across the web app and Slack
• Work on data models and infrastructure that handle very different data types, tuned for fast reads and writes
• Own quality: testing, monitoring, observability, and incident response
• Keep the delivery pipeline healthy: CI/CD, deployments, and infrastructure hygiene

• CS degree, Bachelor's or Master's, or equivalent
• 4+ years shipping production software on a product team where you've owned something. Side projects are a bonus and can compensate for lack of experience
• Strong problem-solving and solid CS fundamentals, which matter to us more than your framework list or years on the clock
• Comfortable with TypeScript/React and Python/Node, or quick to pick them up
• Some hands-on time with CI/CD, deployments, and observability, and up for owning more here
• High agency and happy working on a small, cross-functional team, * Shipped features with LLM or AI components
• Curious about information retrieval or distributed systems - a bonus, not a requirement
• Built integrations with third-party APIs
• Pick up new tools and languages fast

• You get to choose your own machine and peripherals
• Almost unlimited token budget
• Real architecture ownership you wouldn't touch for years at a bigger company
• A new kind of product: AI-native and agent-powered, not a chatbot wrapper
• We work hard and celebrate the wins, small and big
• Offsites a few times a year to work and play together

full_time

Organization Taizen

We're early-stage, so the risk is real: timelines move, and equity is worth what the company becomes, not what it's valued at today. Here's why we think it's worth it. We're two ex-Spotify engineers who've shipped and scaled software used by millions, funded, and building with real customers every week. We move fast and we break things, but we also run SOC2 and ISO 27001-compliant infrastructure, a combination you won't find in most pre-seed teams. There's also a shift worth betting on. Agents can write boilerplate and wire up endpoints faster than any human now. What they can't do is untangle a vague problem, design a data model that holds up under real usage, or debug something failing in a way nobody expected. That's what we're hiring for. After a few years in the job, the thing that sets you apart isn't how fast you type, it's your judgment, and that's exactly what this role leans on.