SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 250,000 - 300,000 / annual
Inworld AI is a research lab and inference provider building real-time AI infrastructure for consumer-facing applications. The company operates at massive scale—serving over 10 trillion LLM tokens per month to hundreds of millions of users across companions, healthcare, fitness, education, media, and more. Inworld has raised $125M+ from top-tier investors including Lightspeed, Kleiner Perkins, Microsoft, Meta, and Founders Fund, and powers experiences for NVIDIA, Microsoft Xbox, Niantic, and others.
You will join as a Senior Machine Learning Engineer focused on inference optimization and real-time serving. This is a full-cycle ownership role: you'll take models from research, containerize and optimize them for production serving, and ensure they run reliably at scale. The work spans model acceleration techniques, high-performance systems engineering, distributed inference infrastructure, and the APIs developers use to integrate these capabilities.
The team values engineers who thrive in ambiguity, learn fast, and obsess over impact. You'll work on unclear problems and be expected to make them clear—designing benchmarks, prototyping solutions, and questioning architecture when you see a better path to solve latency or throughput challenges. Performance, latency, and reliability are first-class product features, not afterthoughts. The culture is flat, fast-moving, and minimal on process theater. Inworld emphasizes in-person collaboration to solve hard problems and foster team culture, with relocation assistance available.
REQUIREMENTS:
- Deep understanding of modern serving frameworks and techniques (vLLM, TRT-LLM, or equivalent)
- Hands-on experience with model acceleration: quantization, distillation, caching strategies, continuous batching, paged attention, speculative decoding
- High-performance systems proficiency: C++, CUDA, Rust, or highly optimized Python; ability to profile code and extract maximum performance from NVIDIA GPUs
- Experience with distributed systems and scaling: Kubernetes, Ray, custom load balancing, multi-GPU/multi-node inference, handling thousands of concurrent connections
- Demonstrated public work: non-trivial systems programming projects, open-source contributions to major inference engines, or deep-dive technical write-ups
- PhD in CS, Physics, Math, or equivalent practical experience building backend or ML systems
- Comfort with ambiguity, bias toward shipped and stable code, and deep curiosity about architectural trade-offs
