Description

The AI/ML Psychology Apprenticeship provides hands-on experience in designing, building, and testing conversational AI experiences in a production environment with end users. You'll collaborate with an accomplished team of researchers from diverse backgrounds to design, build, test, and validate conversational AI experiences that promote healthy relationships and individual wellbeing.  You will work directly with the CEO, AI Lead and Senior Psychologist. The final outcome of the apprenticeship could be a research paper and/or conference presentation.

What you’ll learn

Our unique approach to AI/ML combines the experiences of Amine Allouah, a former Meta AI researcher with expertise in algorithmic design and decision systems, Daniel Fraser, an entrepreneur and innovator with early hands-on experience in enterprise conversational AI and NLP, and Jenna Rieder, a psychology professor with a background in affect, stress, and trauma. By working with the Kinection team, you'll learn the following methods necessary for building a conversational system or conducting AI research:

• NLU Design: Learn how to extract meaning from unstructured text by designing ontologies and knowledge graphs so that a machine can understand written language.
• Synthetic Training Data Generation: Learn how to use AI to generate synthetic training data to create ML models capable of extracting signals and meaning from text.
• Annotation: Learn the best practices for human-in-the-loop design. Gain hands-on experience with annotation guideline design, Uncertainty sampling, and annotations operations.
• Performance Evaluation: Understand the approaches for benchmarking model performance. Learn how model performance impacts user experience.
• ML Model Optimization: Learn various model optimization methods such as Prompt Engineering, Few Shot Training, Fine-tuning, Proxy tuning, and more.

Inspired by the Conversational AI Framework (CAIF) used by leading AI companies and enterprise AI teams, Kinection uses an algorithmic approach to dialogue state management. It leverages natural language understanding (NLU),  agentic workflows and prompt chaining.

Apprenticeship Stages

The work opportunity is envisaged in three stages over the next 4-6 months.

Stage 1: Learn - Work alongside the Kinection Team: attend team meetings to understand our processes and approaches. The goal is to develop an understanding of technical feasibility and transfer our learnings to you. Meetings include: Weekly AI/ML Engineering meeting, Weekly Annotation team meeting, User Research Review every other week.

Stage 2: Build - Design and build conversational experiences that align with your research interests and can potentially become product features. In this phase, you’ll apply techniques learned from Stage 1. Using our underlying system architecture and methodologies, we will help you design the following: UX, signal extraction methods, prompts, and algorithms.

Stage 3: Research & Publish - Design and conduct research as part of a study to validate your designs or to support your own research agenda. We’ve invested in Beta Testing, a user research platform capable of running longitudinal studies in which users engage with a product daily over the course of weeks or months. With this work, you can publish your findings and/or present them at a conference.

‍

To apply, email Jenna Burk at Jennaburk@kinection.ai

‍