Staff Machine Learning Engineer, Video Ingestion & Serving Platform
TwelveLabs10 hours ago
Seoul, Korea, SouthStaff+
Responsibilities
- Design, develop, and operate backend services and platform systems for large-scale video processing, embedding, and AI model serving.
- Improve concurrency, retries, backpressure, and fault recovery for long-running jobs using durable workflow systems such as Temporal.
- Optimize throughput, latency, GPU utilization, and infrastructure costs using load testing, profiling, and operational metrics.
- Design PostgreSQL data models, sharding strategies, and query paths while leading safe, uninterrupted production operations.
- Build infrastructure and CI/CD with Kubernetes, Terraform, and ArgoCD, and improve scalability using Karpenter and KEDA.
- Advance observability and incident response covering metrics, traces, logs, and alerting to improve service reliability.
- Set technical direction and lead production rollouts in collaboration with Backend, ML, Infrastructure, and Product teams.
- Write production code, resolve incidents, establish technical standards, and provide design reviews and mentoring across teams.
Requirements
- At least 5 years of software engineering experience or equivalent capability.
- Experience designing and operating distributed systems involving workflows, queues, and databases in production.
- Strong proficiency in Go or Python, with practical development ability in the other language.
- Experience directly operating services on Kubernetes and cloud infrastructure and automating infrastructure with Terraform or similar IaC tools.
- Experience performing live migrations of databases, storage, or backend systems while considering service availability and data consistency.
- Experience identifying performance bottlenecks and improving throughput or cost through load testing, profiling, and monitoring.
- Experience leading technical decisions across multiple teams and increasing execution effectiveness through design reviews and mentoring.
- Ability to own problems from definition through production operation in a rapidly changing environment.
- Preferred experience with GPU-based inference serving systems such as KServe, vLLM, or Triton.
- Preferred experience with durable workflow engines such as Temporal, Cadence, or Step Functions.
- Preferred experience with sharded or distributed databases such as Aurora Limitless, Citus, or Vitess.
- Preferred experience building and operating observability stacks using Grafana, Mimir, Loki, Alloy, or OpenTelemetry.
- Preferred experience with FFmpeg, video decoding, transcoding, or large-scale media processing pipelines.
- Ability to collaborate effectively in English with global teams.
Benefits
- Hybrid work arrangement with autonomy and collaboration.
- Latest MacBook, KRW 700,000 worth of home-office equipment, and equipment replacement every three years.
- Unlimited LLM tokens for responsible and effective AI use in technical roles.
- KRW 1,400,000 annual professional development allowance for courses, conferences, and memberships.
- English education and global buddy programs.
- Taxi fare support for late-night and weekend commuting.
- Annual KRW 7,200,000 corporate card allowance usable for meals, transportation, and similar expenses.
- Office snack bar with snacks, coffee, and seasonal fruit, plus dinner allowance for office work after 7 p.m.
- Annual health checkups for the employee and one family member.
- Group insurance with a choice of accident, dental, or family accident coverage, plus influenza vaccination support.
- Two-week paid year-end holiday break.
- A three-month probationary period applies to all new hires, with 100% salary paid during the period.
- Global team environment working with B2B customers.
Tech Stack
Categories
About TwelveLabs
TwelveLabs builds video-understanding infrastructure and multimodal AI models that let developers and enterprises search, analyze, and build applications from video and audio. It sells APIs and platform integrations, working with major clouds (AWS, Azure, GCP) and data platforms (Databricks, Snowflake) to support production-scale workloads in media, entertainment, sports, security, and government. Founded in 2021 and headquartered in San Francisco, it is privately held.