Founding ML Infrastructure Engineer

Stellenbeschreibung:

Back to Careers

Founding ML Infrastructure Engineer

Engineering Remote Full-time $160k–$220k + equity

Join Mixpeek as one of the first engineering hires. You'll design and scale the multimodal infrastructure behind our video/image/audio/document indexing and retrieval platform. You'll own Ray-based pipelines, inference services, and the storage layer (Qdrant, MongoDB, ClickHouse, S3) that power our APIs. This is a founding role—you'll shape our core architecture and engineering culture.

Responsibilities

Design, build, and optimize Ray-based pipelines for large-scale multimodal data processing
Own end-to-end flows from object storage (S3) through feature extraction to vector + metadata stores (Qdrant, MongoDB, ClickHouse)
Develop and manage high-performance inference services (e.g., Triton, Ray Serve) behind our public APIs
Implement pragmatic MLOps practices for model and feature extractor deployment, monitoring, and lifecycle management
Work directly with the founder and early customers to debug issues and shape the roadmap
Build monitoring, logging, and alerting for data pipelines and inference services
Continuously improve retrieval performance and relevance across our feature stores and vector databases

Requirements

Senior-level experience (4+ years) building and operating production data/ML or distributed systems
Strong proficiency in Python and experience with ML frameworks (PyTorch, TensorFlow)
Hands-on experience with Ray (Core/Data/Serve) or similar distributed frameworks, and eagerness to go deep
Solid understanding of distributed systems, data-intensive workloads, and performance tuning
Experience with vector databases (ideally Qdrant) and hybrid search over embeddings + metadata
Comfortable working across infra, backend APIs (FastAPI), and a fast-moving product roadmap
Strong problem-solving and communication skills; you enjoy owning systems end-to-end