ЕвропаCantina
Data Scientist (Real-Time Video Generation)
21 ч. назад
ЕвропаSeniorRemote
generative modelsvideo model trainingfine-tuningstreaming generationlatencycontinuitymodel evaluationinference optimization
Cantina is an AI startup developing AI characters and a social AI platform. The role involves research and development of real-time video generation models for interactive characters, focusing on expression, motion and efficient streaming generation.
About the company
- Cantina Labs is a social AI company, developing a suite of advanced real-time models that push the boundaries of expression, personality, and realism.
- We bring characters to life, transforming how people tell stories, connect, and create.
- We build and power ecosystems.
- Cantina, our flagship social AI platform, is just the beginning.
About the product
- We’re working on real-time video generation for interactive characters, spanning realistic humans, stylized characters, and non-human identities.
- A character needs a consistent appearance, expressive and coherent motion, and a natural relationship between speech and movement.
- The model also needs to respond quickly and keep generating reliably as an interaction unfolds.
- Quality, controllability, and efficiency all matter.
Responsibilities
- Develop models supporting responsive interaction while maintaining visual quality and consistency over time.
- Investigate latency, continuity, and model behavior changes during longer generation.
- Explore appearance, facial expression, gesture, and body movement integration into coherent characters.
- Improve speech and motion relationship across identities and visual styles.
- Adapt video models to new data, behaviors, and capabilities through fine-tuning and post-training.
- Investigate training objectives and conditioning methods in implementation and evaluation.
- Make generation faster and more efficient via distillation and inference optimization.
- Work on model-level approaches and runtime improvements measuring effects on quality, latency, and memory.
- Develop training data by identifying missing coverage and failure modes, and improve curation and sampling.
- Build comparisons for visible progress and understandable failures combining quantitative and qualitative evaluation.
- Read relevant research, reproduce and adapt ideas, share experiments, demos, and findings.
- Help the team prioritize efforts based on experimental results.
Requirements
- 3+ years of experience in generative models.
- Experience training or fine-tuning video or related generative models.
- Strong understanding of streaming generation techniques, latency, and continuity.
- Familiarity with data development and evaluation methods.
- Ability to learn new methods quickly and implement useful experiments.
- Good experimental judgment, ownership, and clear communication skills.
Conditions
- Remote (Europe).
- Competitive salary and equity depending on skills, experience, job scope, location, and market data.
- For U.S.-based roles: generous medical, dental, and vision insurance; 42 days paid time off including PTO, sick days, and holidays; parental leave and fertility support.
- 401(k) retirement savings plan.
- Lifestyle spending account of $500/month.
- Complimentary lunch and snacks for in-office employees.
- One Medical membership and more.