Responsibilities:
- Business Understanding and Scope Definition: Work with stakeholders to understand the business problem that the AI model aims to solve. Help define the project scope, translating business requirements into technical specifications. identify relevant data sources and determine key performance indicators (KPIs).
- Data Acquisition and Preprocessing: Work with ML engineers in designing pipelines collecting appropriate data from various sources, cleaning and preprocessing the data, and ensuring data quality.
- Model Selection and Training: Design appropriate training strategies (e.g., supervised learning, reinforcement learning) and appropriate configuring of model parameters. Design and select appropriate ML algorithms and architecture (LLM architecture (e.g., BERT, GPT-3) based on project requirements.
- Evaluation and Optimization: Recommend the metrics and design reports used to evaluate the model’s performance using various metrics, such as accuracy, precision, recall, and F1-score. Identify areas for improvement and optimize the model by adjusting parameters, trying different architectures, or incorporating new data.
- Prompt Engineering and Interaction Design: Designing prompts that effectively communicate with the LLM and elicit the desired responses Phrase prompts to get the best results and avoid unintended consequences.
Experiment with different prompts and evaluate their impact on the LLM's performance.
- Deployment and Monitoring: Work with Engineers to deploy the AI model into a production environment. Recommend the metrics and reports to be used to track model performance. Contribute to the setting up of automated monitoring systems and developing strategies for handling unexpected behaviour.
- Collaboration and Communication: Collaborate with other team members, including ML engineers, product managers, and domain experts. Communicate their findings and recommendations to stakeholders.