Research Intern

Tavus; Greater London; Remote; GBP 17,000 - 28,000; Full time

Benefits offered by this job

Remote internship

Job summary

Tavus, a London-based research lab, invites MSc/PhD students to join as interns on human computing projects.

You'll work with a researcher from day one on challenging, technically demanding topics across multimodal ML, Python/PyTorch, and real-time AI humans.

Internships run 4–6 months, with remote options for EU/US timezones; strong candidates may contribute to publications.

Qualifications

• PhD student with experience in human computing.
• MSc students with strong programming skills in Python/PyTorch.
• Interest in multimodal machine learning.

Responsibilities

• Work with a researcher to own a project from day one.
• Contribute to ongoing research directions with potential for publications.
• Internships run for 4–6 months and may be remote.

Skills

Python

PyTorch

Multimodal ML

Education

PhD student

MSc student

Job description

Tavus https://www.tavus.io/ is a research lab pioneering human computing. We’re building AI Humans — a new interface for natural, face-to-face interaction between people and machines.

Our real-time models let machines see, hear, respond, and even appear human, enabling conversations that feel immediate and lifelike. By combining human-like emotional intelligence with the scale and reliability of software, we’re creating agents that can communicate naturally, 24/7, in any language.

We’re a Series B company backed by Sequoia Capital, Y Combinator, and Scale Venture Partners.

Our research

We push the boundaries of real-time human–AI interaction across audio, video, and multimodality. We aim at multimodal conversations with natural turn-taking, interruptions, and communication with AI-humans that are natural and expressive.

• Conversational modeling
• We explore how to represent and model human communication beyond text. This involves understanding and integrating multiple signals in conversation (including emotions, intonation, intent, personality) and translating them into representations that enable natural interaction via control of models for audio and video generation.
• We work on real-time speech systems focused on generating and transforming human-like audio. Our efforts center on conversational speech generation with high quality, expressiveness, and robustness across a wide range of conditions and speakers.
• We develop systems for generating and rendering realistic human video in real time. This includes ensuring visual consistency, responsiveness, and controllability in interactive settings.
• Optimization
• We focus on enabling real-time performance across the full system. This includes reducing latency and ensuring smooth, continuous interaction between participants.

Our work spans both fundamental modeling and systems challenges, with opportunities to contribute to publishable research.

Internship opportunities (MSc / PhD)

There are several internship opportunities across the four themes.

The internships will run for 4-6 months (preferably 6 months) and you’ll be able to contribute to the team from day 1. You’ll work closely with a researcher and take ownership of a project in collaboration with the team. Projects are designed to be technically challenging and aligned with ongoing research directions, with strong potential for high-impact outcomes (including publications where appropriate).

The ideal candidate would be a PhD student with experience in one or more of the topics above and strong interest in the domain of human computing. Highly motivated MSc students with excellent programming skills in Python/PyTorch and experience in Multimodal ML will be considered as well.

Location

We have offices in London and San Francisco. We are open to remote internships for very strong candidates in EU or US compatible timezones.