Product ML Engineer — Memory-Driven AI Agents

Alpaca Research; Cambridge (MA); Presencial; USD 160.000 - 180.000; Jornada completa

Descripción de la vacante

Alpaca Research is building a full-stack AI platform and is seeking an ML engineer to bridge model capabilities with complete systems. The role focuses on model optimization, memory, agent architectures, and evaluation to deliver reliable, efficient AI that users can own.

You will work across model fine-tuning, memory design, and inference-time techniques, balancing speed and quality on local hardware while integrating ideas into real-world products used by people.

Formación

• Strong ML fundamentals with modern language models, transformers, PyTorch, and training/inference techniques.
• Experience with LLM agents, tool calling, planning, retrieval, or multi-step reasoning systems.
• Understanding of long-context processing, memory architectures, and context window management.
• Familiarity with quantization, efficient inference, and tradeoffs between quality, latency, and compute.
• Experience with supervised fine-tuning, RL, distillation, or related post-training techniques.

Responsabilidades

• Improve models by fine-tuning, distilling, and optimizing for tool use and complex agent tasks.
• Build agent systems including harnesses, tool-use architectures, planning loops, and execution frameworks.
• Engineer memory and context with retrieval and persistent memory systems for long tasks and conversations.
• Optimize inference with quantization and inference-time techniques on local hardware.
• Evaluate and iterate through benchmarks for agent capabilities, reliability, memory, and real-world tasks.
• Research and implement new papers and ideas, prototyping and integrating successful approaches into production.

Conocimientos

ML fundamentals

Agentic systems

Memory & context

Model optimization

Post-training

Research-minded engineering

Independence

Herramientas

PyTorch

Transformers

Descripción del empleo

Alpaca Research is building a full-stack AI platform and is seeking an ML engineer to bridge model capabilities with complete systems. The role focuses on model optimization, memory, agent architectures, and evaluation to deliver reliable, efficient AI that users can own.

You will work across model fine-tuning, memory design, and inference-time techniques, balancing speed and quality on local hardware while integrating ideas into real-world products used by people.