---
title: "Research Scientist, Generative Video & Neural Avatars"
slug: "research-scientist-generative-video-neural-avatars"
id: "bf64dc11-f987-4c3f-891b-c851e6cfd6ca"
department: "AI Research"
department_code: "RES"
location: "London, UK"
workplace_type: "Hybrid"
employment_type: "Full-time"
compensation: "£120k – £170k (annual package)"
is_remote: false
is_featured: false
headcount: 1
canonical_url: "https://jobs.talent247.co/jobs/research-scientist-generative-video-neural-avatars"
markdown_url: "https://jobs.talent247.co/jobs/research-scientist-generative-video-neural-avatars.md"
apply_endpoint: "https://jobs.talent247.co/api/public/jobs/research-scientist-generative-video-neural-avatars/apply"
schema_endpoint: "https://jobs.talent247.co/api/public/jobs/research-scientist-generative-video-neural-avatars/form-schema"
webmcp_tool: "apply_for_role"
updated_at: "2026-11-15T17:28:47.059Z"
content_signal: "search=yes, ai-input=yes, ai-train=no"
---

# Research Scientist, Generative Video & Neural Avatars

## Position Telemetry

- **Reference:** `RESEARCH-SCIENTIST`
- **Practice / Sector:** AI Research (`RES`)
- **Location:** London, UK
- **Workplace Type:** Hybrid
- **Employment Type:** Full-time
- **Compensation:** £120k – £170k (annual package)
- **Canonical HTML URL:** https://jobs.talent247.co/jobs/research-scientist-generative-video-neural-avatars

## Role Dossier & Specification

THE CLIENT & MANDATE
Our client is a $4B generative video pioneer headquartered in London, backed by tier-one global venture capital and strategic AI chipmaker investors, whose synthetic media platform is used by over 90% of the Fortune 100. Within their 40+ person London R&D organization, their Interactive Avatars & Generative Video Research group is pushing the frontier of photorealistic digital humans that move, gesture, emote, listen, and converse indistinguishable from live studio footage.

They have engaged Talent247 to recruit a Research Scientist, Generative Video & Neural Avatars in London. In this high-ownership research role, you will invent and scale next-generation audio- and intent-conditioned video diffusion and flow-matching models capable of generating expressive full-body human motion, micro-expressions, contextual listener reactions, and rock-solid identity preservation over long interactive horizons.

KEY RESPONSIBILITIES
• Conceptualize, prototype, and pre-train large spatiotemporal diffusion and flow-matching transformer architectures conditioned on streaming speech audio, text intent, 3D pose/camera controls, and user interaction cues.
• Solve core scientific challenges in long-horizon interactive avatar streaming: preventing temporal drift, maintaining sub-frame lip-sync accuracy across 140+ languages, and synthesizing natural listener back-channeling (nods, gaze shifts, smiles).
• Explore 3D-aware priors, implicit human representations, and world-model physics so digital presenters interact coherently with scenes, lighting, and objects.
• Lead post-training initiatives—including DPO/RLHF for human perceptual preference, identity/motion disentanglement, and step distillation—to bring research breakthroughs to commercial shipping quality.
• Collaborate closely with capture-studio, multimodal data-curation, and GPU inference teams to co-design high-value training datasets and real-time serving architectures.
• Publish selective, high-impact work and help shape the multi-year research roadmap for interactive multimodal agents.

TECHNICAL ENVIRONMENT & OFFERING
Researchers operate on dedicated multi-node H100/B200 GPU clusters with proprietary studio multi-camera capture datasets and a modern PyTorch distributed training stack. Based hybrid in central London, our client offers £120,000–£170,000 GBP base salary plus substantial Series E equity, comprehensive UK health and family benefits, and full visa/relocation support.

## Core Requirements

ESSENTIAL QUALIFICATIONS
• PhD in Computer Vision, Machine Learning, or Computer Graphics (or equivalent industrial research track record) with first-author publications at CVPR, ICCV, ECCV, SIGGRAPH, NeurIPS, ICML, or ICLR.
• Deep theoretical and hands-on mastery of modern generative visual modeling: video diffusion transformers (DiTs), flow matching, spatiotemporal VAEs/tokenizers, and multi-modal conditioning.
• High proficiency in PyTorch, including multi-node distributed training, custom loss design, mixed-precision optimization, and perceptual video evaluation.
• Track record owning end-to-end research projects from literature synthesis and dataset formulation through large-scale training and ablation studies.
• Strong communication skills and enthusiasm for seeing frontier research used daily by millions of people.

PREFERRED / BONUS EXPERIENCE
• Direct research experience in 3D/2D digital human avatars, audio-driven facial/body animation, talking-head synthesis, or interactive world models.
• Familiarity with causal video streaming, distillation, or joint audio-video generation architectures.

## Application Schema & Direct Apply (`apply_for_role`)

Candidates and authorized AI agents on behalf of a candidate can submit an application via the declarative WebMCP form tool `apply_for_role` on https://jobs.talent247.co/jobs/research-scientist-generative-video-neural-avatars or via HTTP `POST https://jobs.talent247.co/api/public/jobs/research-scientist-generative-video-neural-avatars/apply` (`application/json` or `multipart/form-data`). Non-browser HTTP agents must include header `X-Talent247-Intake-Token` obtained from `GET https://jobs.talent247.co/api/public/intake-token` (or automatically returned in `GET https://jobs.talent247.co/api/public/jobs/research-scientist-generative-video-neural-avatars/form-schema`).

### Form Fields

- `full_name` (text, required): Full name
- `email` (email, required): Email
- `phone` (phone, optional): Phone
- `location` (text, optional): Location
- `linkedin_url` (url, optional): LinkedIn URL
- `cover_letter` (textarea, optional): Cover letter
- `consent` (checkbox, required): I consent to my data being processed for this application
- `cv_file` / `cv_base64` + `cv_filename`: Upload a CV in .pdf or .docx format (<= 10 MB) or provide an `https://` LinkedIn profile URL in `linkedin_url`.

### Alternative: Talent Pool Intake (`join_talent_pool`)

If this mandate is not an exact match, register once in the Talent247 Talent Pool at https://jobs.talent247.co/talent-pool (`https://jobs.talent247.co/talent-pool.md`) via WebMCP tool `join_talent_pool` or `POST https://jobs.talent247.co/api/public/talent-pool`.
