AI Research Scientist

kadence; United States; On-site; USD 180,000 - 250,000; Full time

Benefits offered by this job

Equity pre-funding

Deferred salary until funding

Influence over product and architecture

Job summary

A pioneering AI video startup is looking for a Chief Data Scientist to take on a hands-on role in building the infrastructure for generative video AI. This position requires expertise in large-scale video datasets, Python engineering, and collaboration with ML teams. Join this exciting opportunity to shape the technical future of the company and enjoy significant equity and influence before funding rounds.

Qualifications

• Direct hands-on experience with large-scale video datasets for AI training.
• Experience with leading video generation companies.
• Strong understanding of video structure and data governance.

Responsibilities

• Architect and prototype video data labeling systems.
• Build tools for optimizing video data for AI training.
• Deploy scalable human-in-the-loop and machine-assisted annotation workflows.
• Design provenance tracking and rights verification for datasets and outputs.
• Deploy scalable human-in-the-loop plus machine-assisted annotation workflows.
• Lead QA for labeled data, model performance, and feedback loops.
• Represent Kadence as a technical authority with enterprise customers.

Skills

Large-scale video datasets preparation

Python engineering

Data labeling frameworks

Communication skills

Tools

CVAT

Labelbox

FiftyOne

Custom tools

Job description

Overview

We are an early-stage, revenue-generating startup building the foundational infrastructure for the next generation of AI video. Our platform transforms real-world video and likeness data into AI-ready assets, ensuring that outputs can be used, published, and monetized with proper rights and provenance. We are live with enterprise customers and have a strong pipeline of partnerships. We are seeking a Chief Data Scientist to join us as a co-founder and help shape the technical future of the company.

Role

This is a hands-on builder role, not a pure research position. You’ll be coding, prototyping, and working directly with enterprise customers to extend our platform from large-scale video processing into AI training pipelines and SDKs. You’ll collaborate closely with the founding team to design and implement systems that will power the world’s most advanced generative video platforms.

We’re looking for someone who moves fast, thrives in high-stakes environments, and has the technical credibility to engage directly with ML teams and foundation model developers.

What You’ll Do

• Architect and prototype video data labeling systems that support large-scale ingestion, shot segmentation, and frame-level tagging
• Build tools to optimize video data for AI training, including shot selection, compression, metadata enrichment, and SDK/API integrations
• Extend infrastructure into training pipelines (LoRA injection, ControlNet integration, Nvidia-style SDKs)
• Design provenance tracking, data fingerprinting, and rights verification systems for both datasets and model outputs
• Deploy scalable human-in-the-loop + machine-assisted annotation workflows
• Lead quality assurance systems for labeled data, model performance, and feedback loops
• Represent the company as a technical authority in conversations with enterprise customers and GenAI partners

What We’re Looking For

• Direct, hands-on experience preparing large-scale video datasets for AI training (not just text or images)
• Experience working at or with leading video generation companies (e.g., Scale.ai, OpenAI Sora, Google Veo, Runway, Adobe Firefly, Stability, Kling, Nvidia, etc.)
• Strong Python engineering skills and experience with data labeling frameworks (CVAT, Labelbox, FiftyOne, or custom-built tools)
• Familiarity with LoRA training, ControlNet, diffusion models, or transformer-based video generation workflows
• Deep understanding of video structure: shot continuity, scene labeling, object/character tracking, bounding boxes, cinematic annotations
• Bonus: Knowledge of copyright, data governance, and AI provenance systems
• Strong communication skills and the ability to engage as a peer with ML teams and technical customers

Why Join Kadence

This is a rare ground-floor opportunity to join a company with active revenue, real customers, and patents pending—before institutional funding.

• Equity: Pre-funding, meaning 10x–20x more equity than a post-funding hire
• Salary: Nominal or deferred until funding (expected within 3–5 months), followed by market-rate comp
• Influence: Shape the product, architecture, and team. Your code will power the next wave of generative video AI.
• Timing: We’re at the inflection point—when joining matters most.