Data Scientist | 8+ Years of Experience | Machine Learning Engineering, AWS, Azure, Deep Learning
Actively Looking for Work
Experienced Data Scientist and ML Engineer
Experienced Data Scientist & ML Engineer with 8+ years of designing, developing, and deploying machine learning solutions across the energy and analytics domains. With a PhD in Theoretical Physics and an MSc in Data Science, I’ve built end-to-end pipelines for anomaly detection, forecasting, real-time inference and predictive diagnostics—leveraging modern cloud platforms. Skilled at turning business challenges into robust, scalable ML solutions. Currently advancing expertise in MLOps, LLMs, and full-stack ML deployment.
-
Core Skills & Expertise:
- Programming & Data Tools: Python, SQL, Pandas, NumPy, Scikit-learn, Jupyter, Bash
- Machine Learning & AI: XGBoost, Deep Learning (CNN, LSTM), Time Series Forecasting, NLP, Topic Modeling, Supervised/Unsupervised Learning, Self-Supervised Learning
- Frameworks & Libraries: TensorFlow, PyTorch, Transformers (Hugging Face), FastAPI, MLflow, SHAP
- Data Engineering & Pipelines: ETL design, Feature Engineering, Real-time & Batch Inference, Data Cleaning, Data Leakage Management
- Cloud & DevOps: AWS (S3, Lambda, EC2, SageMaker, ECR), Azure ML, Docker, GitHub Actions, CI/CD workflows
- MLOps & Experimentation: MLflow, Model Registry, Experiment Tracking, Model Monitoring
- Visualization & BI: Matplotlib, Seaborn, Plotly, Tableau
- Edge & Full-stack ML: TensorFlow Lite, Dockerized Inference, REST API deployment, Edge-cloud coordination
Key Projects & Experience
-
Itron Inc., United States – Principal Data Scientist
- Led the design and implementation of one of the company’s core EV detection pipelines using smart meter data as Table Stakes Product. My contribution is as follows:
- Performed Data Exploration, Feature Engineering, Model Exploration, Training and Monitoring: Used deep learning (CNN, LSTM) and XGBoost, Azure ML for training, MLflow for monitoring, Azure Blob Storage for Feature Store.
- Optimized Models by: Hyper Parameter Tuning, Pruning and Quantization achieving business metrics requirements.
- Implemented offline Dockerized Inference deployment to Customer Analytics Dashboard utilizing AWS ECR, AWS S3, FastAPI
- Inventor of the US Patent (App. no. 63/541,686) for the modeling and design of the above pipeline.
- Acted as a technical lead on cross-functional efforts spanning firmware, cloud integration, and product management, ensuring analytics outputs aligned with business requirements.
- Built a generative AI PoC for predictive diagnostics using domain-specific language models.
- Mentored junior team members on best practices in team's ML projects including experimentation, reproducibility and Pythonic development.
-
Xylem Inc., United States – Lead Data Scientist
- Revamped the company’s analytics pipeline for infrastructure maintenance, building a suite of reusable algorithms in Python and SQL over AWS Redshift, to identify sensor anomalies, fault patterns, and provide predictive insights. This pipeline is called Revenue Locator, a company's flagship product.
- Transitioned analytics from ad hoc models to repeatable, automated workflows integrated with SaaS dashboards for client delivery, enhancing operational efficiency.
- Engaged regularly with product and business stakeholders to scope analytics goals, define success criteria, and deliver data-driven recommendations, resulting in an increase in deployment health and performance KPIs for utility customers.
- Mentored interns participating in data science projects resulted in growth and improvement in product development.
-
Whiskerlabs Inc., United States, Data Scientist II
- Developed smart home energy models tailored to household patterns using Python-based ML workflows.
- Built event-driven pipelines using AWS Lambda to process telemetry streams and enable real-time recommendation updates.
-
Recent Work (Independent Projects)
- Developed RAG-based diagnostic assistants using LLMs and FAISS.
- Created smart grid forecasting pipelines with edge/cloud inference integration.
- Hands-on experimentation with LLM fine-tuning and prompt engineering.
-
Education
- PhD in Theoretical Physics – University of Oklahoma, USA
- MSc in Data Science – Galvanize (University of New Haven), San Francisco, CA
Relocation Note
Itron Inc., United States – Principal Data Scientist
- Led the design and implementation of one of the company’s core EV detection pipelines using smart meter data as Table Stakes Product. My contribution is as follows:
- Performed Data Exploration, Feature Engineering, Model Exploration, Training and Monitoring: Used deep learning (CNN, LSTM) and XGBoost, Azure ML for training, MLflow for monitoring, Azure Blob Storage for Feature Store.
- Optimized Models by: Hyper Parameter Tuning, Pruning and Quantization achieving business metrics requirements.
- Implemented offline Dockerized Inference deployment to Customer Analytics Dashboard utilizing AWS ECR, AWS S3, FastAPI
- Inventor of the US Patent (App. no. 63/541,686) for the modeling and design of the above pipeline.
- Acted as a technical lead on cross-functional efforts spanning firmware, cloud integration, and product management, ensuring analytics outputs aligned with business requirements.
- Built a generative AI PoC for predictive diagnostics using domain-specific language models.
- Mentored junior team members on best practices in team's ML projects including experimentation, reproducibility and Pythonic development.
Xylem Inc., United States – Lead Data Scientist
- Revamped the company’s analytics pipeline for infrastructure maintenance, building a suite of reusable algorithms in Python and SQL over AWS Redshift, to identify sensor anomalies, fault patterns, and provide predictive insights. This pipeline is called Revenue Locator, a company's flagship product.
- Transitioned analytics from ad hoc models to repeatable, automated workflows integrated with SaaS dashboards for client delivery, enhancing operational efficiency.
- Engaged regularly with product and business stakeholders to scope analytics goals, define success criteria, and deliver data-driven recommendations, resulting in an increase in deployment health and performance KPIs for utility customers.
- Mentored interns participating in data science projects resulted in growth and improvement in product development.
Whiskerlabs Inc., United States, Data Scientist II
- Developed smart home energy models tailored to household patterns using Python-based ML workflows.
- Built event-driven pipelines using AWS Lambda to process telemetry streams and enable real-time recommendation updates.
Recent Work (Independent Projects)
- Developed RAG-based diagnostic assistants using LLMs and FAISS.
- Created smart grid forecasting pipelines with edge/cloud inference integration.
- Hands-on experimentation with LLM fine-tuning and prompt engineering.
Education
- PhD in Theoretical Physics – University of Oklahoma, USA
- MSc in Data Science – Galvanize (University of New Haven), San Francisco, CA
Currently based in the U.S., I am planning to relocate to the UAE to join relatives and contribute to impactful, data-driven organizations.