--- title: 'Senior Machine Learning Engineer, Pegasus at TwelveLabs' canonical: 'https://feeny.ai/job/senior-machine-learning-engineer-pegasus-twelvelabs-seoul-mh9vx0x3ab12' type: 'job' last_seen: '2026-09-10' --- # Senior Machine Learning Engineer, Pegasus at TwelveLabs - **Company:** TwelveLabs - **Location:** Seoul, South Korea - **Employment:** full-time - **Work type:** hybrid - **Posted:** 2026-08-19 - **Last confirmed live:** 2026-09-10 - **Apply:** https://jobs.ashbyhq.com/twelve-labs/a8fa59d5-e364-43ec-9a5a-dc98a2f3ed53 ## Job description ## Who we are Video is 90% of the world's data. Most of it is invisible to machines. TwelveLabs builds the intelligence layer to change that. Our multimodal AI models understand video the way humans do — across sight, sound, and motion — and power production-scale AI workloads across media, entertainment, sports, security, and government. We have raised more than $210 million from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, Databricks, Index Ventures, NAVER Ventures, Korea Investment Partners, Quadrille Capital, Red Bull Ventures, and AI pioneers including Fei-Fei Li, Silvio Savarese, and Alexandr Wang. We are a global company, headquartered in San Francisco with offices in Seoul, New York, and London, and employees around the world. We believe the differences in our cultural, educational, and life experiences make our products stronger. Building technology that understands the world in all its complexity requires people who see it from every angle. We are looking for individuals who are driven by hard problems and want their work to matter. Come build it with us! ## About Jockey Jockey is TwelveLabs' unified agentic system that reasons across your videos and images. It combines a reasoning model with a memory layer that builds a knowledge store from your corpus. No context window holds a video archive. We work at a million hours of video. A single model forward pass can tell you about one file; it can't reason across a corpus, and no context window closes that gap. Jockey decomposes a query, retrieves, segments, and reasons across thousands of videos and images. Point it at an archive, ask for a highlight reel or the best viral moments, and it returns timestamped cuts you can use. Corpus-level understanding you can act on is the whole product. Built for agents, not just people. As AI agents increasingly become the primary consumers of video, we're building production-grade infrastructure that scales to millions of hours while delivering reliable, high-quality results for both human users and autonomous agents. We build on models we own. Marengo, our embedding model, resolves a query like "the moment we almost missed the flight" into real retrieval. Pegasus, our video-language model, returns structured, timestamped moments on a schema you define. We ship and improve both continuously, so Jockey's quality compounds with every release — no re-integration for customers. Few teams get to build an agent on a stack they control end to end. Deep expertise, one system, open culture. Foundation models, knowledge construction, search, and the agent harness all live in one org. Each team owns its domain and is expected to have deep expertise in it — but like a Formula 1 team, we optimize for the global system, not local parts. A model gain that doesn't expand what the agent can do isn't a gain. We trace a single algorithm change through to end-system behavior, and share work in progress weekly, not just finished results. Anyone can pull the context they need from any team. ## About the team The Cognition Models team owns the models that turn video into structured understanding and reasoning: Pegasus, our video-language model, and Jockey Core, the reasoning LLM behind Jockey. In the model stack we sit between Perception Models (embeddings and retrieval) and the agent system — taking what's retrieved and producing structured understanding and the reasoning to act on it. We focus on multimodal systems with high instruction-following capability and complex, hierarchically structured outputs. Our work spans training infrastructure from pre-training to RL, temporal segmentation and structured metadata extraction, large-scale inference and serving systems, data-curation and evaluation pipelines, and building Jockey Core. We ship products with real-world value rather than doing research in isolation, working as a goal-oriented, cross-functional team of ML researchers and engineers — using the most advanced compute in the world, including NVIDIA B300s, to accelerate the research-to-production cycle. ## About Pegasus Pegasus is TwelveLabs' video-language model — it turns video into useful analysis by reasoning over visuals, speech, audio, and on-screen text. A key capability is Segment, our time-based metadata feature: instead of a broad question about a video, customers define the exact segment types they care about and the metadata fields they want back, and Pegasus finds the relevant start and end times and returns structured metadata for each segment — titles, summaries, topics, people, visual subjects, confidence, or domain-specific labels. This turns video into time-based, structured data that flows directly into search, archive, editing, compliance, or content-management workflows. In this role, you will - Build, improve, and operate production ML systems for Pegasus, with a focus on reliability, performance, and maintainability. - Work across core parts of the ML stack, including deployment, inference, evaluation, monitoring, and supporting infrastructure. - Develop systems for serving Video Language Models (VLMs) and handling multimodal data and metadata at production quality. - Make strong technical decisions within your area and drive execution with a high degree of ownership. - Explore and adopt AI-assisted development tools such as Claude, Gemini, and GPT to improve productivity across coding, experimentation, debugging, and documentation. You may be a good fit if you have - Strong software engineering and machine learning fundamentals. - Experience building and shipping ML systems in production. - Experience with multimodal data and familiarity with areas such as computer vision, natural language processing, LLMs, or VLMs. - Experience with distributed ML or data workflows, ideally in Kubernetes-based environments. - Strong engineering judgment around performance, reliability, and maintainability in production environments. ## Preferred qualifications - Experience serving or optimizing LLM/VLM systems in production. - Experience with inference optimization techniques such as batching, caching, or quantization. - Experience building AI/ML systems from early-stage development through production deployment. - Master's or PhD in Machine Learning, Computer Science, or a related technical field. Read more about the team - [영상에 진심인 곳은 전 세계에 몇 군데 없어요](https://www.twelvelabs.io/ko/blog/kiankim-pegasus-research-interview) - [아무리 뛰어난 모델도, 안 쓰이면 ‘신기하다’에서 끝이에요](https://www.twelvelabs.io/ko/blog/sjkim-mle-interview) - [Pegasus 1.5를 만든 사람들](https://www.twelvelabs.io/ko/blog/pegasus-1-5-seoul-builders) - https://www.twelvelabs.io/ko/blog/%EB%B9%84%EB%94%94%EC%98%A4%EB%A5%BC-%EA%B5%AC%EC%A1%B0%ED%99%94%EB%90%9C-%EC%9E%90%EC%82%B0%EC%9C%BC%EB%A1%9C-time-based-metadata(tbm)-%ED%8C%8C%EC%9D%B4%ED%94%84%EB%9D%BC%EC%9D%B8-%EA%B5%AC%EC%B6%95%EA%B8%B0 - [Cutting Edge Isn’t Plug-and-Play: B300에서 FlashAttention-4 커스터마이징하기](https://www.twelvelabs.io/ko/blog/cutting-edge-isn%E2%80%99t-plug-and-play) ## Benefits and Perks Growth & Tools - 글로벌 B2B 고객과 함께 성장하는 Global Team - 자율성과 협업을 모두 갖춘 하이브리드 근무 - 최신 맥북 및 70만 원 상당 재택근무 장비 지원, 3년 주기로 최신 장비 교체 - Tech 직군이 AI를 효율적이고 책임 있게 활용할 수 있도록 LLM 토큰을 무제한 지원 - 강의, 컨퍼런스, 멤버십 등에 사용 가능한 연 140만원 상당 자기개발비 지원 - 영어 교육 프로그램 및 글로벌 버디 프로그램 운영 - 야간 및 주말 출퇴근 택시비 지원 Meal & Snack - 식비·교통비 등 자유롭게 사용할 수 있는 연 720만원 상당 법인카드 제공 - 사무실 내 스낵바 운영 (간식, 커피, 제철 과일 등) - 사무실 근무 시, 오후 7시 이후 저녁 식대 제공 Wellness & Family - 연 1회 본인 및 가족 1인의 건강검진 제공 - 단체보험 가입 (상해보험/치아보험/가족 상해보험 중 택 1) - 독감 예방접종비 지원 - 연말 2주간 유급 Holiday Break 운영 Additional Notes - 한국 전 직군 공통 자격요건: 해외 여행에 결격 사유가 없는 분 - 모든 입사자에게 3개월의 수습기간이 적용됩니다. 해당 기간 동안 급여는 100% 지급됩니다. - 채용 과정에서 제출한 서류에 허위 사실이 있거나 부정 행위가 확인될 경우, 합격이 취소되거나 향후 채용이 제한될 수 있습니다. ## About TwelveLabs ## Company Overview - **One-liner**: TwelveLabs provides a powerful video intelligence platform that enables enterprises to search, analyze, and generate insights from video content using multimodal, video-native AI models. - **Entity Type**: Private (Series A) - **Headquarters**: San Francisco, California, United States - **Founded**: 2021 - **Founders**: Not publicly available ## Core Business - **Primary Industry**: Software Development / Artificial Intelligence - **Target Customers**: B2B; Enterprise (Media & Sports, Advertising, Government/Public Sector, Security, Creative Industries) - **Mission Statement**: To pioneer the first multimodal, video-native AI that sees and understands like humans do, unlocking the full potential of video for the world. ## Products & Services - **Marengo**: A multimodal embedding model that turns video into searchable data (spatiotemporal embeddings) supporting 47 languages with 78.5% composite accuracy. - **Pegasus**: A video language model that reasons continuously over the full temporal arc of video (up to two hours), enabling entity tracking, causation, and narrative understanding. - **Video Search & Discovery API**: Allows natural language search across entire video libraries to locate specific actions, scenes, dialogue, and emotions. - **Video Understanding Platform**: Infrastructure for ingesting, indexing, and analyzing video at massive scale (up to 60x real-time speed, 10k+ hours per day) via a single API pipeline. ## Market Standing - **Valuation**: Not publicly available - **Total Funding**: $110.1M USD (as of late 2025) - **Annual Revenue**: Approximately $4.5M USD (as of latest estimates) - **Notable Investors**: New Enterprise Associates (NEA), NVentures (Nvidia), Index Ventures, Naver Ventures - **Notable Advisors**: Fei-Fei Li (Professor, Stanford University), Silvio Savarese (Professor, Stanford University), Jeffrey Katzenberg (Former CEO, Dreamworks Animation), Alex Wang (CEO, Scale AI), Lukas Biewald (CEO, Weights & Biases) - **Growth Signals**: Headcount grew 102.1% year-over-year to 149 employees; operates across 6 countries (US, South Korea, Canada, India, UK, Philippines); achieved AWS AI Competency; announced integration with Snowflake AI Data Cloud. ## Competitive Advantages - **Video-Native Architecture**: Models are designed from the ground up for video (unlike general-purpose LLMs), offering superior temporal reasoning and cross-modal understanding (video, image, audio). - **SOTA Accuracy**: Marengo achieves 78.5% composite accuracy in video understanding benchmarks. - **Scale and Speed**: Indexes video at 60x real-time speed, enabling processing of 10,000+ hours of video per day. - **Security and Compliance**: SOC 2 Type II certified with encrypted data handling and flexible deployment options. - **Strong Talent Pool**: Team includes alumni from Meta AI, Nvidia, Cruise, Scale AI, and Cohere, with advisors from top academic and industry institutions. ## Strategic Focus - **Product Development**: Advancing foundation models (Marengo, Pegasus 1.5) for deeper video understanding and reasoning. - **Go-to-Market**: Expanding enterprise adoption across key verticals: government/public sector (surveillance, compliance), media (content archives, highlights), advertising (contextual brand safety), and creative industries. - **Infrastructure & Integration**: Building out MCP (Model Context Protocol) integrations and partnerships with cloud data platforms like Snowflake. - **Global Expansion**: Scaling operations in APAC (Seoul office) and Europe (London office) to capture growing demand. ## Why Work Here - **Culture & Mission**: Described as a place where employees solve “AI’s toughest challenges” in video understanding, building systems that “see, hear, and understand like humans.” Emphasis on curiosity, vision, and deep expertise. - **Team & Environment**: Collaborative environment with a mix of research scientists and engineers. Team began with 12 members from diverse backgrounds in language, video, ML, and perception. - **Work Policy**: Hybrid/office-based with physical locations in San Francisco, New York, Seoul, London, and Pangyo (South Korea). - **Growth & Impact**: Rapidly scaling company (102% headcount growth) with heavy investment in R&D (26% of employees in technical roles, 8% in research). Opportunity to shape foundational video AI technology. - **Notable Perks**: Work alongside leading AI researchers and engineers; opportunity to publish research and participate in the broader AI community. ## Sources 1. [twelvelabs.io (Company Website)](https://www.twelvelabs.io/) 2. [twelvelabs.io (About Us)](https://www.twelvelabs.io/about-us) 3. [twelvelabs.io (Careers)](https://www.twelvelabs.io/careers) 4. [linkedin.com (Company Profile)](https://www.linkedin.com/company/twelvelabs) 5. [jobs.ashbyhq.com (Job Listings)](https://jobs.ashbyhq.com/twelve-labs) ## Other roles at TwelveLabs - [Staff ML Research Engineer, Multimodal Structure & Marengo](https://feeny.ai/job/staff-ml-research-engineer-multimodal-structure-marengo-twelvelabs-seoul-6nsxx2vahw5e) — Seoul, South Korea - [Staff ML Research Engineer, Pegasus - Training Performance](https://feeny.ai/job/staff-ml-research-engineer-pegasus-training-performance-twelvelabs-seoul-2f3nf2dwrzn3) — Seoul, South Korea - [Senior Platform Engineer](https://feeny.ai/job/senior-platform-engineer-twelvelabs-seoul-jzxx921vzdgb) — Seoul, South Korea - [Senior Executive Assistant](https://feeny.ai/job/senior-executive-assistant-twelvelabs-san-francisco-55dq5jshz3jy) — San Francisco, CA - [Head of Applications](https://feeny.ai/job/head-of-applications-twelvelabs-united-states-z0rgf3aywvz5) — United States - [Director of Engineering, Deployment](https://feeny.ai/job/director-of-engineering-deployment-twelvelabs-united-states-ej522ydb97qq) — United States - [Business Systems & Analytics Lead](https://feeny.ai/job/business-systems-analytics-lead-twelvelabs-san-francisco-f4gmw265a6nc) — San Francisco, CA - [Token Engineer](https://feeny.ai/job/token-engineer-twelvelabs-seoul-y5ges4ftqpxp) — Seoul, South Korea - [Tech Lead Manager, Jockey Core](https://feeny.ai/job/tech-lead-manager-jockey-core-twelvelabs-seoul-nr4gsqtgepwt) — Seoul, South Korea - [Senior GTM Systems Engineer](https://feeny.ai/job/senior-gtm-systems-engineer-twelvelabs-united-states-2pnynhf8j7zg) — United States