ML Engineer

Indent: SF_OP_211893-5

Role: ML Engineer

Location: Oakland/MTV/San Diego CA (3 days Work from Office)

Max Rate: $65/hr

Responsibilities

• Design and own shared AI/ML components, frameworks, and pipelines - spanning model training, evaluation, and fine-tuning - used across multiple product teams
• Apply ML fundamentals and LLM techniques (prompting, fine-tuning, RAG) to solve concrete customer problems, then evaluate model performance in production
• Prototype and build agent builder frameworks and embedded AI experiences, taking cutting-edge agentic AI concepts from early exploration to product-ready implementations
• Drive rapid prototyping and experimentation to move from proof-of-concept to high-accuracy, performant systems
• Architect and build full-stack, AI-native applications end-to-end, from backend services to production LLM integrations

Qualifications

• BS or higher in Computer Science or equivalent practical experience
• 5+ years building production AI/ML systems; prior experience leading an engineering effort a plus
• Strong CS fundamentals - data structures, algorithms, system design - plus solid ML fundamentals (classification, regression, clustering, neural networks)
• Proficiency in Python and standard ML frameworks (PyTorch, TensorFlow, Pandas/NumPy)
• Practical experience with LLMs: prompt engineering, fine-tuning, and frameworks like LangChain
• Experience building or operating agentic systems, agent frameworks, or long-term memory/context architectures for AI applications
• Familiarity designing evaluation frameworks or offline/online eval pipelines for measuring model and agent quality
• Track record of launching AI integrations in production and evaluating their real-world impact
• Strong cross-functional collaboration skills, partnering effectively with data scientists, product managers, and engineers across global teams
• Proficiency in one or more full-stack languages (JavaScript, Java) and experience designing scalable services - microservices, relational/NoSQL data stores, Kubernetes