SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Wispr AI is building the voice interface for computing. The company's first products—Flow (natural voice interaction in any application) and Notetaker (context building across conversations)—represent the foundation of a vision to create an interface that can perceive, understand, and take action with earned trust.
As System Architect for New Products, you will design and build core platform primitives that enable the broader engineering team to prototype, test, and scale AI product concepts. You'll own the reference architecture spanning models, live context, memory, tools, agent harnesses, and durable execution—making deliberate decisions about what belongs in each layer and which components should become general platform primitives for other teams to adopt.
Key responsibilities include:
- Build shared platform primitives for agentic systems development, including memory management, context management, tools, and background execution frameworks.
- Create infrastructure for prototyping, training, and evaluating models, with reusable environments and developer tooling.
- Develop systems for triggers and context capture across devices, surfaces, and modalities.
- Design harnesses and primitives for coordination and execution handoff between models.
- Ensure these systems are production-ready through observability, recovery, security, and latency/cost controls.
You'll work across product, ML, client, and infrastructure teams, communicating architecture decisions and tradeoffs clearly while driving decisions to closure. The role demands starting from user experience requirements (latency, quality, privacy, trust) and working backward to system design.
Wispr is a talent-dense team that holds strong opinions, tests them quickly, and builds technology that sparks joy. The company is venture-backed and focused on building the first voice interface used daily by a billion people.
REQUIREMENTS:
- Demonstrated experience designing and shipping significant greenfield systems from ambiguous problem definition through production use, with evolution based on real failures and usage.
- Track record building shared primitives adopted by multiple teams, with deliberate choices about abstraction boundaries and developer experience.
- Ownership of backend or distributed systems with critical scaling and reliability requirements.
- Ability to start design from user experience and its latency, quality, privacy, and trust requirements.
- Proven ability to create alignment across product, ML, client, and infrastructure teams; communicate architecture and tradeoffs clearly; and drive decisions to closure.
- Exceptional depth in backend and platform systems, with ability to learn adjacent areas (ML, inference, agentic systems) quickly.
- Architecture-level ownership demonstrated in prior roles.
Note: You do not need to have trained foundation models, built a low-level inference stack, or worked in every domain mentioned. The company sponsors H1B, O1, EB1, L1, STEM OPT, and other visa categories and will make reasonable efforts if an offer is extended.
