Job description

About Inworld Inworld is a research lab and inference provider focused on realtime AI for consumer-facing applications. We build first-party speech models, serve LLMs, and run the inference behind modular APIs designed for high-volume, realtime workloads. Hundreds of millions of users interact with Inworld powered apps every day and we serve over 10 trillion LLM tokens per month. Our models and infrastructure support consumer applications across companions, healthcare, fitness, education, media, and more. Our work spans model research, realtime inference, large-scale serving infrastructure, and the APIs developers use to bring these capabilities into production. We’ve raised more than $125M from Lightspeed Venture Partners, Section 32, Kleiner Perkins, Microsoft’s M12 venture fund, Founders Fund, Meta, Stanford, and others. Our technology has powered experiences from companies including NVIDIA, Microsoft Xbox, Niantic, Logitech Streamlabs, Wishroll, Little Umbrella, and Bible Chat. Inworld has also been recognized by CB Insights as one of the 100 most promising AI companies globally and named one of LinkedIn’s Top 10 Startups in the USA. ABOUT THE ROLE Inworld Inference inworld.ai serves the models consumer apps run on. We're growing it from an API into a full platform: one place where a team can go from its first API call to its own model, tuned on its own traffic and running on capacity reserved for it. As a Staff or Principal Software Engineer, you'll design that platform's core architecture and write much of its code. You'll define the abstractions every customer workload runs on, and own the system in production at tens of trillions of tokens a month. WHY THIS ROLE - A platform, not a feature. You'll shape the foundation that every future Inference product builds on. The architecture decisions are yours to make. - Real scale from day one. Realtime inference for hundreds of millions of end users, with latency budgets measured in milliseconds. - Hard problems. Scheduling scarce GPUs across many tenants and workloads, model lifecycles from weights to serving, and latency and uptime commitments customers pay for. - Principal-level scope. You'll partner directly with research, product and leadership on where the platform goes next. WHAT YOU'LL DO - Design and build the platform's core services: control plane, model and deployment lifecycle APIs, tenant isolation, metering and quotas. - Own how GPU capacity is allocated and scheduled across shared and reserved workloads, with clear latency, cost and reliability trade-offs. - Build the path that takes a customer's model from weights to production serving, and keeps it healthy as it changes. - Write production code, mostly in Go, every week. You ship services; you don't just configure them. - Define the APIs and abstractions that customers and internal teams build on. Make the right path the easy path. - Lead design reviews and write the technical direction for the platform. Raise the bar for system design across engineering. - Run what you build: set SLOs, take part in incidents, and turn postmortems into design changes. WHAT WE'RE LOOKING FOR - 8+ years building backend or distributed systems in Go, Rust, Java or C++. - You've designed and shipped a multi-service system end to end, and can explain its trade-offs and what broke in production. - Deep understanding of distributed systems: consistency, failure modes, queuing, load balancing and capacity modelling. - Experience building a platform or developer product with external customers: cloud compute, PaaS, serverless, or an API platform. - Clear technical writing. You can turn an ambiguous problem into a design doc that others build from. NICE TO HAVE - Schedulers or resource managers for scarce hardware (GPU clusters, batch systems, cloud compute). - LLM inference serving (vLLM, SGLang, TensorRT-LLM) or ML training infrastructure. - Multi-tenant systems with usage-based billing or contractual SLAs. - Comfortable with Kubernetes, Terraform and GitOps (ArgoCD). IS THIS ROLE FOR YOU? This is a software engineering role. If most of your recent work is maintaining CI/CD pipelines and infrastructure-as-code, this probably isn't the fit. Candidates must be based in the SF Bay Area or willing to relocate (you will be working on-site in our South Bay office a few days a week). The US base salary range for this full-time position is $280,000 - $350,000. In addition to base pay, total compensation includes equity and benefits. Within the range, individual pay is determined by work location, level, and additional factors, including competencies, experience, and business needs. The base pay range is subject to change and may be modified in the future. Inworld Jobs Privacy inworld.ai

Requirements

Must-haves

A platform, not a feature. You'll shape the foundation that every future Inference product builds on. The architecture decisions are yours to make.
Real scale from day one. Realtime inference for hundreds of millions of end users, with latency budgets measured in milliseconds.
Hard problems. Scheduling scarce GPUs across many tenants and workloads, model lifecycles from weights to serving, and latency and uptime commitments customers pay for.
Principal-level scope. You'll partner directly with research, product and leadership on where the platform goes next.
Design and build the platform's core services: control plane, model and deployment lifecycle APIs, tenant isolation, metering and quotas.
Own how GPU capacity is allocated and scheduled across shared and reserved workloads, with clear latency, cost and reliability trade-offs.
Build the path that takes a customer's model from weights to production serving, and keeps it healthy as it changes.
Write production code, mostly in Go, every week. You ship services; you don't just configure them.
Define the APIs and abstractions that customers and internal teams build on. Make the right path the easy path.
Lead design reviews and write the technical direction for the platform. Raise the bar for system design across engineering.
Run what you build: set SLOs, take part in incidents, and turn postmortems into design changes.
8+ years building backend or distributed systems in Go, Rust, Java or C++.
You've designed and shipped a multi-service system end to end, and can explain its trade-offs and what broke in production.
Deep understanding of distributed systems: consistency, failure modes, queuing, load balancing and capacity modelling.
Experience building a platform or developer product with external customers: cloud compute, PaaS, serverless, or an API platform.
Clear technical writing. You can turn an ambiguous problem into a design doc that others build from.

Nice-to-haves

Schedulers or resource managers for scarce hardware (GPU clusters, batch systems, cloud compute).
LLM inference serving (vLLM, SGLang, TensorRT-LLM) or ML training infrastructure.
Multi-tenant systems with usage-based billing or contractual SLAs.
Comfortable with Kubernetes, Terraform and GitOps (ArgoCD).
IS THIS ROLE FOR YOU?
This is a software engineering role. If most of your recent work is maintaining CI/CD pipelines and infrastructure-as-code, this probably isn't the fit.
Candidates must be based in the SF Bay Area or willing to relocate (you will be working on-site in our South Bay office a few days a week).

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aillmapigorustjavamlkubernetesci/cd

Platform: ashby
Effort: quick