Machine Learning Engineer

Aitopics - San Francisco, CA

Full-time - Mid Level

Onsite - San Francisco, CA

About the position

As a Machine Learning Engineer at Supernormal, you will be responsible for developing AI solutions that enhance the core product experience for meetings, including transcription, note generation, and task automation. This role involves leading the full development cycle of AI solutions, utilizing state-of-the-art techniques to improve meeting assistant scenarios, and collaborating with product and design teams to create new NLP-driven product experiences.

Responsibilities

• Lead the full development cycle of AI solutions for meeting notes, question answering, agent conversations, and task completion.
• Prompt engineering using state-of-the-art techniques to improve core meeting assistant scenarios.
• Build and ship machine learning models to improve transcript quality and reduce API token usage.
• Train and deploy custom language models, fine-tuning them for diverse business needs.
• Create new NLP & LLM-driven product experiences that improve with user feedback.
• Improve LLM-powered search and question answering using retrieval augmented generation (RAG).
• Advocate for and build new and better ways of doing things.

Requirements

• 3-5 years of production-level AI/ML experience building machine learning systems.
• 1+ year experience building NLP systems using LLMs.
• Solid engineering background with a robust foundation in software engineering principles.
• Bachelor's degree in Computer Science, Engineering, AI, Mathematics, or related field.
• Proficient in Python, with experience interfacing with Ruby on Rails and Typescript.

Nice-to-haves

• Experience with LoRA, RLHF, and instruction-tuning for training models.
• Familiarity with optimizing infrastructure for AI solutions.

Benefits

• Full healthcare coverage (Medical, Dental, and Vision)
• WFH budget to ensure you have everything needed to do your best work.
• Education credit (up to $500 per year)
• Fully remote work environment with no return-to-office requirements.