MALAN EVANS

As Lead AI Engineer – Customer-Facing AI Betting Assistant,

• Led development of Sportsbet’s AI betting assistant, enhancing search functionality by enabling natural language queries to find and build bets.
• Joined the project post-POC and delivered the MVP by transforming an experimental AI prototype into a production-grade, customer-facing application, implementing best engineering practices including error handling, version control, and unit/integration testing. Successfully released to 10% of customers.
• Built an evaluation framework combining offline and online methods, leveraging LLM-as-a-judge and Python-based function evaluations. Offline evaluations reduced human review effort by 80%.
• Assumed technical leadership, guiding the team to improve response quality and completeness by 50%, while increasing bet returns by 300% over time.
• Championed A/B testing of multiple approaches, including agentic workflow architectures, to refine performance and user experience.
• Collaborated with internal and external stakeholders on feature design and experimentation with tools, frameworks, and LLM providers.

Other Contributions,

• Introduced AI platform capabilities by developing a Python package that streamlined AI application development and enforced approved usage patterns.
• Mentored junior and senior ML Engineers on leveraging ML platform features such as model explainability, drift detection, and feature stores.
• Initiated training of small language models (SLMs) for classification tasks such as gambling issue detection in user queries. Fine-tuned transformer models using PEFT with LoRA (work in progress).
• Submitted initial proposals for a MCP server for Sportsbook.

• Deployed and maintained NLP-based text analytics services in production environments.
• Designed, implemented, and maintained scalable ML model training and evaluation pipelines.
• Collaborated closely with Data Scientists to set up ML-based feature experimentation workflows.
• Developed lightweight prototype applications and wrote solution previews for new features.
• Published ML-related Event Carried State Transfer (ECST) data to the data lake to support service improvement.
• Designed and developed a Machine Learning Platform to support integration of LLM-based features across product teams.
• Released a new engagement survey comment topic model based on fine tuned transformer models, improving topic classification accuracy by 35%.
• Contributed to the successful release of an LLM-based AI comment summarisation feature for early access customers within 3 months.
• Led the development of a data annotation application to collect high-quality human-labeled data for ML experiments.
• Championed ML model monitoring and LLM application monitoring to identify model drift and LLM response issues in production.

• Led developing end-to-end ML projects for improving internal DevOps task efficiency.
• Developed a model-driven tool that predicts potential regression test failures for Synopsys customers. This tool significantly enhanced the electronic design regression testing turnaround time by 80%.
• Automated feature engineering and ML model training pipelines that ensure model promotion to production only upon meeting predefined evaluation metrics.
• Demonstrated strong customer focus by leading the team to align efforts with client needs and satisfaction.
• Built and maintained robust ETL pipelines to process and transmit diverse customer datasets efficiently.
• Served as the primary contact for the data analytics and machine learning team, facilitating communication and collaboration.
• Mentored, coached, and trained team members to enhance their technical and professional development.
• Conducted internal training sessions and tutorials on data science and software engineering topics to upskill team members.
• Carried out extensive experimentation to improve ML model performance using a variety of libraries and frameworks.