- 🔭 Engaged In: Dissertation research on detecting sepsis in neonates using predictive modeling and data analysis.
- 👯 Collaboration Call: Currently working on advanced machine learning and MLOps pipelines for real-time data predictions.
- 🤝 Support Wanted: Exploring time series forecasting and optimization techniques in production systems.
- 📚 Sharing Insights: Passionate about teaching and sharing knowledge in machine learning and data science.
- Advanced Variational Autoencoders (VAEs) and generative models
- Best practices for scalable MLOps in cloud environments
- Machine Learning models, MLOps pipelines, and data-driven decision-making
- Via Email: [email protected]
- By night, I transform from a data scientist into Gotham's silent guardian. Yes, I am Batman!
- Developed a fault-tolerant data streaming pipeline using Kafka, Faust, and Bash scripting with Zookeeper for reliability.
- Built, registered, and deployed machine learning models using Databricks and MLflow, containerized via Azure Container Registry (ACR), and orchestrated on AKS with GitHub Actions and Terraform.
- Streamed real-time data from EventHub to deployed models, ensuring low-latency predictions.
- Deployed Prometheus and Grafana for monitoring system metrics and visualizing real-time performance insights.
- Developed ML and GenAI models using PacBio RNA isoform data, integrating CI/CD workflows with GitHub Actions and AKS.
- Used Azure Data Factory for orchestrating raw user inputs and storing processed data in Azure Data Lake.
- Designed scalable solutions leveraging AKS, EventHub, and Data Lake for efficient data processing and predictions.
- Integrated monitoring pipelines to track predictions and ensure reliability.
- Implemented a machine learning pipeline for article recommendations using SVD and NMF techniques.
- Productionized the recommendation engine with Streamlit for real-time interaction and improved user engagement.
- Title: "Intelligent Monitoring and Early Disease Detection of Patients in NICUs"
- Developed ML models to predict neonatal sepsis using ECG and vital sign data.
- Engineered features from both tabular (vital signs) and transformed ECG data (GAF, TMF).
- Designed custom architectures to improve predictive accuracy and enable real-time sepsis monitoring.
- Equitably.ai: Developed enterprise-grade AI-powered pricing models leveraging advanced ensemble techniques and time series analytics for revenue growth.
- Platingnum: Built a proof of concept for an IoT project to track foot movements in videos, utilizing transfer learning for precise tracking and analysis.
Python, SQL, Spark, Docker, Kubernetes, MLflow, Azure, Databricks, Prometheus, Grafana