Founding Machine Learning Engineer
$150,000–$300,000 year
RemoteSan Francisco, California, United States
Job Summary
Own the end-to-end development of core ML systems from research and modeling to production deployment. Design and train models for information retrieval, entity resolution, classification, and structured data extraction. Build systems that transform messy, multilingual web-scale data into structured, queryable intelligence. Develop embedding models, ranking systems, and retrieval pipelines for high-precision search and matching. Apply transformer architectures and modern NLP techniques to real-world data problems, leveraging LLMs for extraction, classification, and data enrichment at scale. Continuously evaluate and improve model performance using rigorous experimentation and metrics. Work closely with engineering and product teams to integrate ML systems into production APIs.
Required Qualifications
- 3+ years of experience building and shipping production ML systems, particularly in NLP, information retrieval, or entity resolution
- Strong hands-on experience with Python and PyTorch
- Deep understanding of transformer architectures, including training and fine-tuning encoder models
- Experience building retrieval systems, classifiers, or embedding-based systems
- Experience applying LLMs to structured data problems (e.g., extraction, classification, generation)
- Strong problem-solving skills with the ability to work on ambiguous, large-scale data challenges
- High ownership mindset with a strong bias toward execution in fast-paced environments
Desired Qualifications
- Experience with entity resolution or record linkage at scale
- Background in multilingual or cross-lingual NLP
- Experience building taxonomies, ontologies, or knowledge systems
- Familiarity with distributed training on GPU clusters
- Experience scaling LLM inference pipelines in production
- Research publications or open-source contributions in NLP/IR
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.