Platform Engineer, Infrastructure
No longer in the current catalog. Last included 2026-09-09. Check the employer’s posting for availability.
Technologies
TypeScript · React · Python · LLM · speech models · voice interface
About Wispr Flow
Builds Flow, an AI voice-dictation app that turns speech into formatted, context-aware text across IDEs, Slack, email and docs, monetized via consumer/pro and team subscriptions.
Series A · 50–200 people
Job description
The full responsibilities and requirements are on the employer’s site.
Open application page ↗Source and classification
Internal deployment & tooling · Evidence for this classification:
500ms around the globe. But latency is just the beginning. Our platform layer sits behind every personalization feature, the multi-step workflows that turn speech into action, and the primitive infrastructure the rest of the product is built on. We're shipping new surfaces fast, which means the core platform now needs far more rigor and a lot more headroom. We've built for millions of users and thousands of enterprises with a three-person platform team. You'd be on the ground floor, creating the systems that support the next hundred million. What You'll Do Architect and own whole parts of the platform end to end. Take the gnarly production problems nobody else has picked up. Make every product engineer faster — dev experience, deploy speed, abstractions. Build the agentic workflows that speed up you and your team. This role is perfect for you if: You've rewritten an entire
More from the job description
About Wispr We're building the first voice interface a billion people use every day. Our first product, Flow, lets you talk instead of type anywhere on your computer or phone, and turns messy speech into polished text that is ready to send. It's scaled from $2M → $50M in ARR in under 12 months with millions of users and thousands of enterprise customers. How we think → The team is still small enough that every person shapes what gets built. About the Role Every time someone talks to Flow, they're waiting — mid-sentence, cursor blinking — for their words to appear. 500ms feels like magic. 1000ms feels annoying. People use Flow hundreds of times a day, and they feel every single request. Why this is hard: we've scaled our user base faster than our infrastructure could keep up with. We build and deploy our own voice models, and we've tuned our networking stack to hit P50 latencies of 500ms around the globe. But latency is just the beginning. Our platform layer sits behind every personalization feature, the multi-step workflows that turn speech into action, and the primitive infrastructure the rest of the product is built on. We're shipping new surfaces fast, which means the core platform now needs far more rigor and a lot more headroom. We've built for millions of users and thousands of enterprises with a three-person platform team. You'd be on the ground floor, creating th [... source excerpt omitted ...] ast, but never without rigorous testing. You get excited about complex problems with lots of constraints and edge cases. You care as much about the user journey as the technical requirements. Tech stack: Python/FastAPI, PostgreSQL, Redis, Temporal/SQS, Terraform, AWS Logistics Location. About 60% of the team is in San Francisco and the rest are remote, mostly US-based. If you're in SF: you'll be in the office most days. We think the best work happens in person. Most of the team is here five days a week, and a few do three or four days depending on family or long commutes. If you're remote: we're open to remote for some roles. We fly remote teammates to SF for three to four day
Employer postings · Data from · Sources