Furqan M.

-Audit Logging & Platform Compliance- Supporting efforts to streamline audit log capture within Palantir Foundry (Life Sciences) by reviewing existing logging behavior against compliance requirements. Co-developing test strategy using Playwright to validate reliable and compliant audit logging. -AI/ML Lifecycle Governance & Validation- Supporting audits and validation of Foundry’s AI/ML lifecycle workflows—including data ingestion, training, deployment, and monitoring—against 21 CFR Part 11 and other life sciences regulatory requirements.

As part of the UC Davis MSBA: - Dispenser Analysis Chatbot: Designed and deployed a GPT-4 powered dispenser-analysis chatbot that interprets telemetry data and surfaces actionable insights, reducing troubleshooting time and improving technician productivity. - Module Replacement Prediction Model: Designed and orchestrated end-to-end ML pipelines for a predictive maintenance system that forecasts Flow Control Module (FCM) failures; automated feature engineering, training, validation, and deployment using cloud-native tooling. - Built automated, scalable data pipelines in Python and SQL to ingest, clean, and transform dispenser telemetry data for model ingestion, training, and evaluation - Led a cross-functional team (DS, SWE, BA) to prototype, validate, and communicate ML solutions, ensuring alignment with business objectives and stakeholder adoption.

- Regulatory Health Authority Chatbot: Performed data exploration and feature engineering for a regulatory RAG chatbot using LLaMA3/Mistral, enabling scientists to anticipate health authority questions from prior submissions and reducing preparation time by 30%. - Flywheel Vertex AI Integration: Conducted technical analysis of ML deployment workflows across GitLab, Flywheel, and Vertex AI, identifying an approach that eliminated 6+ months of development effort for the ML engineering team. - ML Platform Savings Dashboard: Led the development of a user-friendly Metabase dashboard by extracting raw ML platform usage metadata, transforming it into analysis-ready tables in Redshift, and visualizing cost trends, driving over $14K in annual savings. - MLOPS Tools POCs: Evaluated various MLOPS tools focusing on Feature Store and Experiment Tracking/Model Registry capabilities to provide recommendations to stakeholders for our MLOPS ecosystem needs. Tools evaluated included Neptune AI, MLFlow, HopsWorks, and Microsoft PowerBI. - CI/CD SonarQube/Sysdig Integration: Integrated security/code format checks into our CI/CD pipeline using SonarQube. These checks saved data scientists hours of debugging time in the algorithm development process. - Optimized CI/CD pipelines via Python Automation Scripts, resulting in increased efficiency and faster project delivery. - MLTest Lakera + Roche Integration: Integrated Lakera’s MLTest framework/library on Roche servers, enabling Data Science users to visualize model outputs better. This improved model robustness by enabling developers to identify blind spots and make more effective improvements.

- Worked with the Advanced Analytics team to help bring different software such as AWS Sagemaker, Kubeflow Pipelines, and Vertex AI as a Machine Learning service to the company platform called Apollo. - Closely collaborated with Data Scientists to prototype ML Pipelines on various ML platforms such as AWS SageMaker, Vertex AI, Kubeflow, streamlining scientists’ workflow. - Built an end-to-end ML pipeline predicting patient outcomes using clinical datasets via Logistic Regression; containerized workflows with Docker for reproducible deployment.

- JIRA Implementation Project: Worked in collaboration with the IT team on the JIRA System implementation project. Assisted in configuring system workflows, issue fields, and email notification settings for a successful end-user release. Provided end-user support by troubleshooting reported issues. - Microsoft Team Project: Worked to provide new user training including a companywide training session, account setup, and researching to troubleshoot end-user issues. - Saved company $45,000 with my services provided and described above.