Anees Abbasi

At Artifeel, I am working on improving the battery performance of our smart security devices using data-driven approaches. My role involves analyzing device and sensor data, building machine learning models, and experimenting with optimization techniques to extend battery life. I work with tools like AzureML for training and deployment, GitHub for version control, and use Notion and Slack for documentation and team collaboration. It’s a great opportunity to apply data science to real-world hardware challenges in a deeptech environment.

Built ML pipelines using Python, PyTorch, and Scikit-learn, reducing training time by 35%, and developed BERT-based NLP models boosting accuracy by 18%. Designed ETL workflows with Airflow, AWS Lambda, and S3, improving data processing and cutting failures by 40%. Enhanced model transparency with SHAP/LIME and created feature stores using Snowflake and dbt, reducing retrieval latency by 25%. Implemented MLflow/Grafana monitoring and A/B testing, improving model stability by 20%, alongside EDA that increased personalization by 12%. Streamlined deployments with GitHub Actions and Docker (30% faster) and contributed to ML-driven ESG features improving onboarding and data transparency.

Transformed raw telecom data into interactive Power BI dashboards, helping leadership monitor network performance, customer trends, and key service KPIs. Developed DAX-based KPIs (ARPU, churn) to enhance decision-making across customer care and technical teams. Integrated data from SQL Server, Excel, SharePoint, and CRM systems into unified models for improved accuracy and consistency. Presented insights to supervisors and analysts using data storytelling, highlighting trends, anomalies, and actionable recommendations.