---
title: "Senior Research Engineer, Generative World Models"
slug: "senior-research-engineer-generative-world-models"
id: "6e6ad41d-56bc-47b7-a7a8-f7f9aa2d7668"
department: "AI Research"
department_code: "RES"
location: "London, UK"
workplace_type: "Hybrid"
employment_type: "Full-time"
compensation: "£140k – £195k (annual package)"
is_remote: false
is_featured: false
headcount: 2
canonical_url: "https://jobs.talent247.co/jobs/senior-research-engineer-generative-world-models"
markdown_url: "https://jobs.talent247.co/jobs/senior-research-engineer-generative-world-models.md"
apply_endpoint: "https://jobs.talent247.co/api/public/jobs/senior-research-engineer-generative-world-models/apply"
schema_endpoint: "https://jobs.talent247.co/api/public/jobs/senior-research-engineer-generative-world-models/form-schema"
webmcp_tool: "apply_for_role"
updated_at: "2026-11-15T17:31:45.832Z"
content_signal: "search=yes, ai-input=yes, ai-train=no"
---

# Senior Research Engineer, Generative World Models

## Position Telemetry

- **Reference:** `SENIOR-RESEARCH-EN`
- **Practice / Sector:** AI Research (`RES`)
- **Location:** London, UK
- **Workplace Type:** Hybrid
- **Employment Type:** Full-time
- **Compensation:** £140k – £195k (annual package)
- **Headcount:** 2 openings
- **Canonical HTML URL:** https://jobs.talent247.co/jobs/senior-research-engineer-generative-world-models

## Role Dossier & Specification

THE CLIENT & MANDATE
Our client is a globally recognized frontier AI research organization headquartered in London, uniting fundamental machine learning scientists and large-scale systems engineers to solve general-purpose intelligence. Their generative world-modeling division builds real-time, interactive spatiotemporal foundation models that simulate physical dynamics, 3D geometry, and causal agent interactions from raw multimodal streams.

They have partnered with Talent247 to hire a Senior Research Engineer for their Generative World Models team in London. Working directly alongside principal research scientists, the post-holder will architect causal video tokenizers, long-horizon memory mechanisms, and high-throughput multi-node training pipelines that bring interactive world simulators to real-time frame rates.

KEY RESPONSIBILITIES
• Co-design and scale autoregressive, masked-generative, and flow-matching world models capable of maintaining 3D spatial and physical consistency over multi-minute interactive rollouts.
• Engineer compact spatiotemporal latent tokenizers and continuous-discrete variational autoencoders optimized for temporal compression and high visual fidelity.
• Build deterministic, fault-tolerant petabyte-scale video and multimodal action-conditioned data pipelines feeding multi-slice accelerator training jobs.
• Optimize long-context spatiotemporal attention kernels, KV-cache eviction strategies, and speculative frame decoding to achieve interactive sub-50ms step latencies.
• Construct automated physics-consistency, object-permanence, and counterfactual action-controllability evaluation suites.
• Drive end-to-end experimental velocity from clean single-node ablations through full-scale foundation model training runs.

TECHNICAL ENVIRONMENT & OFFERING
Stack: Python, JAX, PyTorch, XLA, custom CUDA/Pallas kernels, distributed petabyte video loaders, multi-pod TPU/GPU supercomputing clusters. Workplace: Hybrid in London, UK. Compensation: £140,000 – £195,000 base salary plus competitive equity/bonus package, comprehensive UK benefits, and full visa and relocation support.

## Core Requirements

ESSENTIAL QUALIFICATIONS
• MS or PhD in Computer Science, Machine Learning, Computer Vision, or equivalent Senior-level research engineering experience in generative modeling.
• Demonstrated production or frontier-lab experience training large-scale video diffusion, flow-matching, or autoregressive multimodal foundation models.
• Mastery of JAX or PyTorch across multi-host data, tensor, sequence, and pipeline parallelism regimes.
• Strong software engineering discipline in Python (and C++ where performance-critical), including reproducible configuration, profiling, and automated testing.
• Deep familiarity with 3D spatial representations, camera geometry, optical flow, or latent video compression architectures.

PREFERRED / BONUS EXPERIENCE
• Published research at CVPR, ICCV, ECCV, NeurIPS, ICML, or ICLR on video generation, world models, neural rendering, or embodied simulation.
• Hands-on experience distilling multi-step diffusion models into single-step or few-step real-time causal generators (consistency distillation, adversarial diffusion distillation).
• Familiarity with coupling learned world models to downstream reinforcement learning or planning agents.

## 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/senior-research-engineer-generative-world-models or via HTTP `POST https://jobs.talent247.co/api/public/jobs/senior-research-engineer-generative-world-models/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/senior-research-engineer-generative-world-models/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`.
