Senior Machine Learning Engineer | MLOps Specialist | Azure & AWS Architect
🧠 Building scalable, secure AI systems from hypothesis to production
I’m a data alchemist on a mission to transform messy reality into reliable insights. With over 5 years of experience across healthcare, finance, and compliance, I specialize in:
- ✅ Production-grade ML Systems with full observability, privacy-awareness, and rollback safety.
- 🌐 Cloud-Native MLOps Pipelines using Azure, Databricks, and AWS (SageMaker, Glue, EventHub, ADF).
- 🧩 Research-to-Deployment Workflows with advanced time series models and clinical applications.
- ⚙️ GenAI + Infrastructure Automation, building pipelines that talk, adapt, and recover.
By night, I decode ECG signals to detect sepsis. By day, I deploy GenAI workflows with drift monitoring, audit trails, and infrastructure-as-code.
- 🌡 Detecting neonatal sepsis in real-time NICU settings with custom DL models and biosignal processing (GAF/TMF).
- 📈 Streaming 10,000+ financial signals for a real estate investment anomaly detector on AWS SageMaker.
- ⚖️ Deploying a GenAI compliance tool using Claude (via Bedrock) to flag internal policy violations.
- 🌪 Building real-time MLOps platforms with AKS, MLflow, and Prometheus for full-stack insight.
- Cloud: Azure (ADF, AKS, ACR, Monitor), AWS (Glue, Lambda, S3, Redshift, SageMaker, Bedrock)
- ML & MLOps: PyTorch, XGBoost, MLflow, Kubeflow, Terraform, Azure ML, Prometheus, Grafana
- Data Engineering: Spark, SQL, Kafka, EventHub, Databricks, Qdrant, Feature Store (Feast)
- Monitoring & CI/CD: GitHub Actions, Azure DevOps, AWS CodePipeline, Jenkins
- Built an end-to-end ML system to detect early signs of neonatal sepsis using biosignal transformation (GAF, TMF).
- Deployed on AKS with Databricks and EventHub for live inference + Grafana dashboards for clinicians.
- Integrated AUC-ROC, calibration metrics, and GDPR compliance for hospital partners.
- Kafka + Azure ML + Databricks = real-time ML inference engine.
- Built a fault-tolerant MLOps pipeline with version control, rollback, and streaming model deployment.
- Monitoring built with Prometheus + Grafana, complete with drift alerts.
- Powered by Claude (Anthropic) + Bedrock for semantic analysis of internal comms.
- Logged all LLM interactions with secure audit trails and business-rule tagging.
- 10K+ portfolio metrics parsed via AWS Glue + SageMaker + Lambda.
- Used Isolation Forest + SHAP for explainable alerts on rent spikes and vacancy gaps.
- Generative AI in enterprise: Prompt tuning, vector search (Qdrant), and context window optimization.
- Forecasting with VAEs & probabilistic models.
- Domain-specific LLM fine-tuning with retrieval pipelines on Azure.
In the data realm, I'm part Iron Man (automation), part Doctor Strange (predictive modeling), and part Alfred (quietly making everything run smoothly). And yes, I moonlight as Gotham's silent guardian. 🦇
- 🧪 GitHub
- 🧠 Equitably.ai Team