Platform Software Architect, Senior

Positron AI is hiring platform software architects to define the hardware/software contract for Asimov and the generations that follow. You will work across silicon architecture, firmware, operating systems, runtime, simulation, security, performance, and developer tooling.

This is a direct technical-leadership role reporting to the Director of Accelerator Platform Software. It is intentionally not buried inside a delivery team: you will maintain a cross-generation view, make durable interface decisions, and still contribute reference implementations where code is the clearest way to resolve ambiguity.

Architecture and Interface Definition

• Define the multi-generation platform software architecture for current Asimov silicon and at least one materially new future architecture, specifying stable, versioned interfaces for boot, commands, memory, DMA, C2C, scheduling, telemetry, errors, reset, security, and host/device interaction.

Cross-System Integration

• Connect GOLD, the Asimov interpreter and runtime, libasimov, firmware, Linux, simulation, and post-silicon tools into one coherent architecture.

Technical Leadership and Execution

• Lead architecture reviews, document decisions and trade-offs, and drive cross-team issues to closure without becoming a process bottleneck.

Prototyping and Mentorship

• Build prototypes or reference code for the riskiest interfaces, and mentor senior engineers across the organization.

Security and Performance Partnership

• Partner with Product Security and Performance Systems so that protection, observability, debuggability, and throughput are treated as first-class architectural properties.

• 10+ years of experience in low-level platform software, firmware, operating systems, runtimes, device drivers, compilers, or adjacent systems, with architecture-level ownership.
• Deep understanding of computer architecture, memory systems, concurrency, device I/O, hardware/software interfaces, and system bring-up.
• Strong C or C++ skills, paired with the judgment to choose when a specification, model, prototype, or production implementation is the right artifact.
• Proven ability to influence multiple engineering teams without relying on reporting authority.
• Demonstrated ability to turn ambiguous product and silicon goals into precise interfaces, milestones, and testable system behavior.

• Experience with AI accelerators, GPUs, networking or storage ASICs, or other high-performance devices.
• Hands-on experience spanning pre-silicon models, emulation, first silicon, and production software.
• Familiarity with trusted boot, model protection, distributed inference, KV-cache systems, or performance modeling.

While this role is currently posted at a specific level, we are a growth-oriented organization and are open to hiring at a more senior level for the right candidate. Please note that this job description serves as a focused but generalized overview of the role; specific responsibilities and impact expectations will be tailored to the experience and seniority of the final hire.

In the first 12 months, the end-to-end Asimov software architecture and ownership map are clear, reviewable, and tied to implementation milestones. High-risk hardware/software contracts are specified and exercised in Terminus or Titan before silicon. A forward-looking architecture track is active for a non-evolutionary future device, while current-silicon delivery remains unblocked.

• Shape the foundational hardware/software architecture for a new generation of AI inference silicon, with direct influence on decisions that will span multiple future product generations.
• Take on a technical-leadership role…