Design and build AI-powered features for VSTAR and other EDI platforms, including intelligent tutoring capabilities, semantic search, content recommendations, and LLM-based tools for learners and educators
Apply appropriate AI implementation patterns and strategies such as RAG architectures, agentic workflows, prompt engineering strategies, and LLM orchestration patterns appropriate to educational use cases
Develop backend services and APIs that expose AI capabilities for integration into VSTAR and other applications, working with the development team to determine appropriate integration patterns
Evaluate vender versus open-source AI products and services based on performance, cost, and reliability considerations
Ensure responsible AI practices, including appropriate guardrails, content filtering, and transparency in AI-assisted features
Build and maintain ML pipelines in Databricks for feature engineering, model training, and evaluation
Deploy models and AI services to production with appropriate monitoring, logging, and error handling
Implement MLOps practices proportionate to our maturity: version control, testing, documentation, and reproducibility
Ensure performance, reliability, and scalability of AI-powered services
Own the full lifecycle of deployed AI features, including maintenance, iteration, and retirement
Partner with data engineering to ensure AI systems integrate cleanly with our data infrastructure
Collaborate with software developers to integrate AI features into existing applications
Proactively communicate progress, challenges, and decisions to the team through regular check-ins, documentation, and asynchronous updates
Work with product and educational leadership to identify high-impact AI opportunities
Contribute to EDI's AI strategy and help establish best practices for responsible AI development in medical education
Maintain clear documentation and support knowledge sharing across the team
Stay current with developments in AI tooling, particularly as they apply to education and knowledge work
Requirements
5 – 7 years of experience is required.
Experience in applied machine learning, AI engineering, or a related field (3+ years) is necessary.
Strong Python skills and experience with ML frameworks such as scikit-learn, PyTorch, or TensorFlow (3+ years) is necessary.
Hands-on experience building applications with LLMs, including prompt engineering, embeddings, retrieval-augmented generation, and agents (1+ years) is necessary.
Experience developing backend services (FastAPI, Flask, or similar) and RESTful APIs (1+ years) is necessary.
Track record of deploying AI or ML features to production environments (1+ years) is necessary.
Comfort with SQL and working with data pipelines (3+ years) is necessary.
Ability to communicate technical concepts clearly to non-technical audiences (3+ years) is necessary.
Experience with Databricks and Azure cloud services (1+ years) is preferred.
Familiarity with MLOps tools and practices (MLflow, model registries, CI/CD for ML) (1+ years) is preferred.
Experience with vector databases (Pinecone, Weaviate, Chroma, or similar) (1+ years) is preferred.
Experience working with multiple LLM providers or open source LLMs and evaluating tradeoffs (1+ years) is preferred.
Background in building predictive models (classification, regression, forecasting) (1+ years) is preferred.
Experience in education, healthcare, or other mission-driven sectors (1+ years) is preferred.
Familiarity with the unique considerations of AI in educational contexts (pedagogical alignment, learner privacy, appropriate automation) (1+ years) is preferred.
Demonstrated self-direction and ownership mentality in previous roles is necessary.