Research Engineer, Reinforcement Learning

Job not on LinkedIn

🔥 0 minutes ago

Apply Now

Find Similar Remote Jobs

Set Up Email Alert For New Jobs

📊 Check your resume score for this job

Improve your chances of getting an interview by checking your resume score before you apply.

Upload Resume

LiveKit

WebsiteLinkedInAll Job Openings

11 - 50 employees

Founded 2020

🔌 API

🤖 Artificial Intelligence

📡 Telecommunications

💰 Venture Round on 2022-09

API • Artificial Intelligence • Telecommunications

LiveKit is a platform designed to facilitate real-time AI and multimedia communications. It enables the instantaneous transmission of audio and video between language models (LLMs) and users, creating real-time multimodal experiences. LiveKit provides tools for developing voice AI, robotics, and livestreaming applications, with ultra-low latency and high resilience. It supports integration with telephony networks, simulating human conversation with advanced features like interruption handling and turn-taking cue detection. The platform is geared towards developers, offering SDK support across all major platforms, allowing applications to be deployed and scaled efficiently. LiveKit is open-source under Apache 2. 0 and offers robust security measures, including end-to-end encryption and compliance with GDPR, SOC2 Type 2, CCPA, and HIPAA.

📋 Description

• Build the environments and verifiers models train against • Own the synthetic data pipeline from generation through quality gates • Run end-to-end training experiments and explain model changes • Build evaluations that releases must pass • Choose and adapt open-weight base models for LiveKit’s tasks • Ensure trained behavior works for voice and text agents • Ship models into production and continuously improve them using real usage • Build post-training capabilities for agents operating across voice, SMS, and chat channels, including long-horizon context and reliable tool use

🎯 Requirements

• Strong Python engineering skills • Experience carrying a model from raw data through production • Ability to treat data as the product, including coverage, diversity, and leakage • Ability to design against models exploiting weak rewards • Comfortable working with GPUs and aware of their limitations • Ability to determine when to train and when not to train • Comfortable working collaboratively in a remote environment • Experience with post-training, fine-tuning, reward design, or reinforcement learning such as GRPO • Experience with RL and fine-tuning frameworks such as TRL, verl, OpenRLHF, or a self-written training loop • Experience with fast rollouts using vLLM or SGLang • Experience with multi-GPU training using FSDP • Experience training tool-using or multi-turn agents • Experience with execution sandboxes, verifiers, evaluation harnesses, or tooling • Experience with open-weight model families such as Qwen or Llama • Experience with LoRA or similar techniques • Legal authorization to work in the US or country of residence • Must indicate whether future sponsorship support is required

🏖️ Benefits

• Equity package • Health benefits • Dental benefits • Vision benefits • Flexible vacation policy • Opportunity to shape the brand of a fast-growing developer platform • Collaboration with a small, senior team that values craft and creativity