Research Engineer: Real-Time Voice ML & Production Systems

Sesame; Bellevue (WA); On-site; USD 190,000 - 320,000; Full time

Benefits offered by this job

401(k) max match

Health, vision & dental benefits

Unlimited PTO"

Flexible spending account (FSA)

Stock options

EAP

Job summary

Sesame is building a future where computers are lifelike, with voice agents that feel natural. We’re seeking an ML engineer to own evaluation pipelines, data tooling, and deployment of real-time voice models powering Sesame’s companion experience.

You will scale training/inference for large models, design robust metrics, and ship end-to-end solutions in a fast-paced startup environment. Collaboration with research, infra, and product is essential; strong autonomy is valued.

Qualifications

• Expert-level PyTorch experience.
• Proven software engineer capable of writing production code.
• Hands-on experience training or fine-tuning large language or other large-scale models.
• Experience designing metrics and harnesses that predict user happiness.
• Deep knowledge of the ML lifecycle: dataset ops, training pipelines, eval frameworks, deployment, and monitoring.
• History shipping complex projects to production—especially user-facing ML systems.
• High agency and the judgment to sprint solo or with a squad.
• Track record of setting technical direction and partnering with product, infra, and research.

Responsibilities

• Own evaluation pipelines and automate offline and live evals for production accuracy.
• Harness data to build safe, versioned, privacy‑aware dataset curation tooling.
• Ship models by prototyping, training, and deploying state‑of‑the‑art voice models.
• Scale training and inference for large workloads with latency, throughput, and cost optimization.
• Wire up monitoring and live evals to surface quality regressions.
• Move ideas from whiteboard to production in days with test dashboards.

Skills

PyTorch

ML engineering

Production code

LLM training

Evaluation metrics

ML lifecycle

Production deployment

Cross-functional collaboration

Job description

Sesame is building a future where computers are lifelike, with voice agents that feel natural. We’re seeking an ML engineer to own evaluation pipelines, data tooling, and deployment of real-time voice models powering Sesame’s companion experience.

You will scale training/inference for large models, design robust metrics, and ship end-to-end solutions in a fast-paced startup environment. Collaboration with research, infra, and product is essential; strong autonomy is valued.