ECCV 2026 Career Opportunities

About Synthesia Synthesia is the world's leading AI video platform for enterprise, trusted by 90% of the Fortune 100. Founded in 2017 and headquartered in London, we've raised $530M+ from top-tier investors (Accel, Nvidia Ventures, Kleiner Perkins, GV) and reached a $4B valuation. We're building products that enhance visual communication and enterprise learning at scale.

The Role

Join our ML Platform team as a Principal Engineer to build and operate the systems that power model training, serving, and deployment at scale. You'll work across infrastructure, backend systems, and tooling—shaping technical direction and making pragmatic architectural tradeoffs as our platform evolves.

This is a hands-on IC role with significant ownership and influence on how we scale models, workloads, and teams relying on it.

What You'll Do

• Design and improve systems supporting model training, evaluation, and production serving
• Build infrastructure and tooling for reliable, scalable, cost-efficient ML workloads
• Develop tools and workflows operable by both humans and agents
• Improve scheduling, monitoring, and debugging on GPUs and cloud infrastructure
• Drive improvements in observability, automation, reliability, and developer experience
• Collaborate with researchers and product engineers to turn pain points into platform capabilities
• Contribute to technical direction and shape platform architecture as we grow

You'll Thrive Here If You Have

• Strong production systems experience (reliability, scalability, maintainability)
• Systems mindset: you think in bottlenecks, failure modes, interfaces, and resource usage
• Solid hands-on cloud infrastructure and Linux experience
• Kubernetes and distributed workload expertise
• Strong coding skills (Python or similar backend languages)
• Sound judgment on where automation drives leverage
• Experience building internal platforms or developer tooling
• Pragmatism: you solve the right problem well, not over-engineer
• Track record of technical leadership and architectural influence

Particularly Relevant - ML infrastructure or model serving in production - Research or data-intensive workload support - GPU-based or performance-sensitive systems - Distributed systems observability and debugging - Terraform, Datadog, GitHub Actions, or similar tools

Bonus Points - Agentic or LLM-powered internal tools - Workflow orchestration (Temporal, etc.) - Research-to-production engineering - Performance optimization and resource allocation

About Synthesia Synthesia is the world's leading AI video platform for enterprise, trusted by 90% of the Fortune 100. Founded in 2017 and headquartered in London, we've raised $530M+ from top-tier investors (Accel, Nvidia Ventures, Kleiner Perkins, GV) and reached a $4B valuation. We're building products that enhance visual communication and enterprise learning at scale.

The Role

Join our Post-Training Team to make foundational models production-ready for text-to-video synthetic humans. You'll work on applied research that directly impacts solutions used by 55,000+ businesses worldwide, adapting generative models and optimizing them for real-world avatar generation.

This is your chance to leverage expertise in post-training diffusion models and push the boundaries of what's possible in generative AI.

What You'll Do

• Apply DPO (direct preference optimization) to pre-trained diffusion models
• Adapt models to extend capabilities through new conditioning inputs
• Build solutions for dubbing and evaluate lip-sync quality
• Implement post-training optimization techniques (quantization, pruning, distillation) for avatar generation
• Analyze and address model performance challenges, ensuring high-quality avatar output
• Stay updated with latest research in diffusion models, adversarial networks, and post-training optimization

You'll Thrive Here If You Have

• Computer Vision or Computer Science background with 3+ years industry experience
• Deep knowledge of recent post-training techniques (distillation, adversarial networks, efficient attention)
• Hands-on experience with generative models for images and/or videos (Diffusion/GAN), preferably in avatar domain
• Proficiency with modern ML and deep learning frameworks
• Strong Python skills and commitment to clean, maintainable code
• Experience with Git and preferably CI/CD workflows
• Mindset: you're interested in research, trying new things, and pushing boundaries

Particularly Relevant Experience - Post-training optimization for diffusion models - Avatar-centric or human-centric generative work - Production deployment of generative models - Experience bridging research and product engineering

Benefits Competitive compensation (salary + stock) · Hybrid work (London, Amsterdam, Zurich, Munich) or remote EU · 25 days annual leave + public holidays · Strong culture with regular team socials

About Synthesia Synthesia is the world's leading AI video platform for enterprise, trusted by 90% of the Fortune 100. Founded in 2017 and headquartered in London, we've raised $530M+ from top-tier investors (Accel, Nvidia Ventures, Kleiner Perkins, GV) and reached a $4B valuation. We're building products that enhance visual communication and enterprise learning at scale.

The Role

Join our Video Pre-Training team to build the next generation of production-grade foundation models for human-centric video generation. This is applied research with direct product impact—your work will power realistic, controllable, and emotionally expressive synthetic humans used by tens of thousands of businesses worldwide.

You'll own end-to-end research and engineering projects from hypothesis to production, working at the intersection of large-scale generative modeling, distributed systems, and production engineering.

What You'll Do

• Develop and scale latent video diffusion models tailored for human-centric video generation
• Design conditioning mechanisms to improve control (pose, emotion, script, camera) without sacrificing fidelity
• Advance distributed training strategies (DDP, FSDP, DeepSpeed, sequence parallelism) under real compute constraints
• Improve training stability at multi-node scale
• Design rigorous evaluation frameworks combining automated metrics and human evaluation
• Optimize inference for low latency, high resolution, and cost efficiency
• Run controlled ablations and experiments to drive high-signal modeling decisions
• Move fast, run multiple hypotheses in parallel, identify signal early, and focus on outcomes

You'll Thrive Here If You Have

• Strong experience training deep learning models at scale
• Strong Python and PyTorch skills
• Hands-on experience with diffusion models (image required; video preferred)
• Large scale multi-GPU and multi-node training experience
• Good understanding of distributed training (DDP, FSDP, DeepSpeed, or similar)
• Ability to design controlled experiments and interpret noisy results
• Research-driven but outcome-focused mindset: you care about shipping, not just publishing
• Clear communicator who presents results scientifically

Nice-to-Have - Video diffusion model experience - Avatar or human-centric generation background - Familiarity with world or interactive models - GANs or VAEs experience - Production inference optimization experience

Tech Stack Python · PyTorch · CUDA · DeepSpeed · Sequence parallelism · AWS · SLURM · Docker · GitHub CI/CD

Benefits Competitive compensation (salary + stock + bonus) · Fully remote from Europe or hybrid (London, Amsterdam, Zurich, Munich) · 25 days annual leave + public holidays · Strong culture with regular team socials

About Synthesia Synthesia is the world's leading AI video platform for enterprise, trusted by 90% of the Fortune 100. Founded in 2017 and headquartered in London, we've raised $530M+ from top-tier investors (Accel, Nvidia Ventures, Kleiner Perkins, GV) and reached a $4B valuation. We're building products that enhance visual communication and enterprise learning at scale.

The Role

Join our Interactive Avatars Team within R&D to work on cutting-edge avatar-centric video diffusion models. You'll own applied research projects that directly impact solutions used by 60,000+ businesses worldwide. This is a hands-on role turning breakthrough ideas into real product capabilities—you'll ship models that make a meaningful impact.

What You'll Do

• Adapt diffusion models to incorporate diverse conditioning signals (audio, motion, interaction cues)
• Develop methods for streaming infinitely long video sequences at real-time rates
• Build the perceptual layer of interactive agents: understand user audio and generate appropriate contextual reactions
• Improve lip-sync accuracy, motion realism, and overall visual quality in video diffusion models
• Build robust evaluation frameworks and test suites for continuous quality tracking
• Collaborate with data teams to define needs and ensure high-quality datasets
• Stay up to date with research in world models, interactive agents, and diffusion models

You'll Thrive Here If You Have

• Strong ML background (diffusion, GANs, VAEs) with relevant industry experience
• Hands-on experience with diffusion models, ideally avatar-centric or video-focused, and up to date with recent advances
• Proficiency in PyTorch and modern ML frameworks
• Strong Python engineering skills, confident with git and version control
• Outcome-driven mindset: motivated to push state-of-the-art research into product impact
• Clear communicator of hypotheses, experiments, and results
• Comfortable owning full project execution from conception to delivery

Particularly Relevant Experience - Audio-conditioned video diffusion models and recent video DiT architectures - Owned full model development pipeline end-to-end: data preparation through training and evaluation - Strong publication record in world models, interactive agents, or video diffusion - Collaborative work with product and data teams to define research direction

Benefits Competitive compensation (salary + stock + bonus) · Hybrid work in London, Amsterdam, Zurich, Munich, or remote EU · 25 days annual leave + public holidays · Strong company culture with regular team socials

About Synthesia Synthesia is the world's leading AI video platform for enterprise, trusted by 90% of the Fortune 100. Founded in 2017 and headquartered in London, we've raised $530M+ from top-tier investors (Accel, Nvidia Ventures, Kleiner Perkins, GV) and reached a $4B valuation. We're building products that enhance visual communication and enterprise learning at scale.

The Role

Join our ML Platform team to build and operate the systems that power model training, serving, and deployment at scale. You'll work across infrastructure, backend systems, and tooling—making our ML platform more reliable, scalable, and agent-friendly as we grow.

This is a hands-on IC role with significant ownership. You'll shape platform evolution as we scale models, workloads, and teams relying on it.

What You'll Do

• Design and improve systems supporting model training, evaluation, and production serving
• Build infrastructure and tooling for reliable, scalable, cost-efficient ML workloads
• Develop tools and workflows operable by both humans and agents
• Improve scheduling, monitoring, and debugging on GPUs and cloud infrastructure
• Drive improvements in observability, automation, reliability, and developer experience
• Collaborate with researchers and product engineers to turn pain points into platform capabilities

You'll Thrive Here If You Have

• Strong production systems experience (reliability, scalability, maintainability)
• Systems mindset: you think in bottlenecks, failure modes, interfaces, and resource usage
• Solid hands-on cloud infrastructure and Linux experience
• Kubernetes and distributed workload expertise
• Strong coding skills (Python or similar backend languages)
• Sound judgment on where automation drives leverage
• Experience building internal platforms or developer tooling
• Pragmatism: you solve the right problem well, not over-engineer

Particularly Relevant - ML infrastructure or model serving in production - Research or data-intensive workload support - GPU-based or performance-sensitive systems - Distributed systems observability and debugging - Terraform, Datadog, GitHub Actions, or similar tools

Bonus Points - Agentic or LLM-powered internal tools - Workflow orchestration (Temporal, etc.) - Research-to-production engineering - Performance optimization and resource allocation

About Synthesia Synthesia is the world's leading AI video platform for enterprise, trusted by 90% of the Fortune 100. Founded in 2017 and headquartered in London, we've raised $530M+ from top-tier investors including Accel, Nvidia Ventures, Kleiner Perkins, and GV. We've reached a $4B valuation and are building products that enhance visual communication and enterprise learning at scale.

The Role

Join our Data team to build the world's best human-centric data lake. You'll work at the intersection of applied research, data engineering, and ML infrastructure—managing the complete lifecycle of data that powers our models. With over a million hours of video and audio at your fingertips, you'll collaborate closely with model training teams to extract features and annotations that elevate dataset quality and performance.

This role sits between traditional research and data engineering, focusing on enhancing model performance through high-quality data rather than architecture alone.

What You'll Do

• Manage end-to-end data lifecycle: sourcing, processing, and delivery for researchers and model teams
• Collaborate with model training teams to understand requirements and translate them into data improvements
• Extract new features and annotations that enhance dataset quality and model performance
• Design and operate large-scale data processing pipelines and workflow orchestration systems
• Drive data curation, labeling strategies, and quality evaluation frameworks
• Enhance infrastructure while influencing longer-term team strategy

You'll Thrive Here If You Have

• Strong background in data-centric, applied Machine Learning with hands-on experience improving model performance through data quality, curation, and labeling
• Experience working on the data layer of Generative AI products, particularly images, video, or audio
• Excellent Python skills with focus on writing clean, maintainable, well-tested code
• Hands-on experience designing, building, and operating workflow orchestration systems
• Proven ability to build and operate large-scale data processing pipelines
• Passion for enhancing model performance through high-quality datasets

Particularly Relevant Experience - Data engineering for video or audio generative models - Large-scale data annotation and quality frameworks - Production data infrastructure and pipeline optimization - Cross-functional collaboration with research and product teams

Benefits Competitive compensation (salary + stock + bonus) · Hybrid work (London, Amsterdam, Zurich, Munich) or remote EU · 25 days annual leave + public holidays · Strong culture with regular team socials

About Synthesia Synthesia is the world's leading AI video platform for enterprise, trusted by 90% of the Fortune 100. Founded in 2017 and headquartered in London, we've raised $530M+ from top-tier investors (Accel, Nvidia Ventures, Kleiner Perkins, GV) and reached a $4B valuation. We're building products that enhance visual communication and enterprise learning at scale.

The Role

Video generation is core to everything we ship at Synthesia. We're looking for a Principal Research Engineer (L7) to own the full technical stack for offline video generation—bridging pre-training and post-training, setting long-term direction, and staying personally present at the hardest parts of the work.

You'll partner directly with research leadership to define strategy, resolve cross-cutting technical problems, and accelerate how research reaches product. This is a senior IC role with outsized scope and influence. You need hands-on experience training large generative models from scratch, not supervising from a distance—you've debugged it, shipped it, and you're genuinely driven to push what's possible in video generation.

What You'll Do

• Own end-to-end technical direction for offline video generation, spanning pre-training and post-training
• Partner with research leadership to define unified roadmaps and drive execution
• Identify and unblock critical technical gaps across the pipeline (architecture, training stability, alignment, coordination)
• Accelerate research-to-production velocity: improve handoffs, increase visibility, and ship faster
• Coach and elevate junior researchers and engineers toward senior technical thinking
• Shape team structure and processes to enable high-velocity, cohesive execution

You'll Thrive Here If You Have

• Proven track record training large-scale video generation models from scratch at millions of hours of data
• Deep experience with post-training techniques at scale (RLHF, GRPO, DPO) and judgment to apply them
• Strong data quality instincts and willingness to question it rigorously
• Ability to think

About Synthesia Synthesia is the world's leading AI video platform for enterprise, trusted by 90% of the Fortune 100. Founded in 2017 and headquartered in London, we've raised $530M+ from top-tier investors (Accel, Nvidia Ventures, Kleiner Perkins, GV) and reached a $4B valuation. We're building products that enhance visual communication and enterprise learning at scale.

The Role

Join our Audio Post-Training Team to develop high-quality, expressive, and real-time synthetic voices used by 60,000+ businesses worldwide. You'll work on applied research turning cutting-edge speech generation research into production-grade systems—from streaming speech synthesis to novel architectures that enhance realism and responsiveness.

What You'll Do

• Develop and evaluate streaming and speech-to-speech systems enabling low-latency, interactive voice synthesis
• Adapt models for new conditioning inputs (emotion, speed, prosody, speaker control)
• Implement post-training optimization techniques (quantization, pruning, distillation) for real-time efficiency
• Integrate and test novel architectures (neural codecs, diffusion, flow-matching) to enhance realism
• Define new evaluation metrics for conversational speech, including latency-aware and online MOS prediction
• Apply DPO and distillation to fine-tune large-scale speech models
• Stay updated with latest research in audio diffusion, autoregressive models, and multimodal LLMs

You'll Thrive Here If You Have

• Strong understanding of generative modeling, ideally applied to sequential or multimodal data
• Hands-on experience with large language models or transformer-based architectures
• High proficiency in PyTorch, including distributed training and model optimization
• Solid grasp of time-series modeling and tokenization, preferably in audio or speech
• Demonstrated ability to prototype quickly, test hypotheses, and iterate efficiently
• Proven end-to-end deep learning experience: data preparation through evaluation
• Strong software engineering skills for contributing to large shared research infrastructure

Particularly Relevant Experience - Real-time or streaming architecture experience - State-of-the-art audio/speech generation architectures (diffusion, neural codecs, flow-matching, autoregressive decoders) - Speech-to-speech or text-to-speech (TTS) systems - Original research contributions: publications or open-source work in top-tier venues (ICASSP, Interspeech, NeurIPS, ICML)

Benefits Competitive compensation (salary + stock + bonus) · Fully remote from Europe or hybrid (London, Amsterdam, Zurich, Munich) · 25 days annual leave + public holidays · Strong culture with regular team socials