Production-Ready Conversational AI Research Engineer

Neura Market; San Francisco (CA); On-site; USD 200,000 - 400,000; Full time

Benefits offered by this job

Vacation policy

Medical benefits

Dental benefits

Vision benefits

Life Insurance

Disability benefits

Retirement plan

Parental leave

Fertility benefits

Lunch in office

Job summary

Decagon is seeking a Research Engineer to build industry-leading conversational AI models that power our agent, taking ideas to production. You will own multi-quarter initiatives that push the agent's reliability, capability, and efficiency forward, and design frontier training, evaluation, and orchestration approaches across the system.

We expect strong engineers who own their work end-to-end, ship real improvements, and make high-impact technical decisions.

Qualifications

• 2+ years of experience in AI/ML engineering or research.
• Prior experience post-training and deploying LLMs in production environments.
• Fluency in Python and modern ML tooling (training, evaluation, data pipelines)
• Track record of taking research ideas from prototype -> reliable, measurable production impact

Responsibilities

• Lead research and engineering efforts to improve core conversational capabilities in production, including instruction following, retrieval, memory, and long-horizon task completion
• Build and iterate on end-to-end models and pipelines that optimize for quality, efficiency, and user experience
• Partner with platform and product engineers to integrate new models into production systems
• Break down ambiguous research ideas into clear, iterative milestones and roadmaps.

Skills

AI/ML engineering

Post-training deployment

Python

ML tooling

Tools

ML tooling

Job description

Decagon is seeking a Research Engineer to build industry-leading conversational AI models that power our agent, taking ideas to production. You will own multi-quarter initiatives that push the agent's reliability, capability, and efficiency forward, and design frontier training, evaluation, and orchestration approaches across the system.

We expect strong engineers who own their work end-to-end, ship real improvements, and make high-impact technical decisions.