Data Analyst
Siemens AG | Munich, Germany
- Analyzed complex datasets to extract actionable business insights and support strategic decision-making.
- Developed and maintained automated dashboards to monitor key performance indicators across multiple departments.
September 2025 - Present
Werkstudent Data Science
Siemens AG | Erlangen, Germany
- Unlocked $10M+ in revenue recovery by modeling order intake gaps with regression and clustering, aligning forecasting insights with enterprise financial goals.
- Orchestrated Python ETL pipelines for 25k+ SAP/SieSales bid records, enhancing data quality through schema validation and Z-score/IQR outlier detection.
- Slashed bid accuracy variance by 22% via automated outlier detection in Plotly dashboards, driving actionable sales strategy insights.
February 2025 - August 2025
Master's Thesis on Demand Forecasting
Siemens AG | Munich, Germany
- Designed a Multitask Learning Model (MTL) (Python/TensorFlow) to forecast hiring demand, capturing regime shifts and reducing forecast errors by 70% (MAE) vs. ARIMA baselines.
- Implemented a Python/Snowflake pipeline analyzing 2.5M+ HR records and headcount trends to prioritize top 100 department ID/role combinations.
- Achieved 92% precision in high-risk talent gaps by integrating Optuna-optimized hyperparameters with MTL, enabling proactive retention strategies.
May 2024 - February 2025
Werkstudent Data Science
Siemens Healthineers AG | Erlangen, Germany
- Accelerated real-time decision-making by architecting Kafka pipelines to replicate 5+ data sources, reducing latency from 24hrs to <1hr.
- Engineered Python ETL for 5TB+ data into Snowflake/MSSQL, improving query speed by 40% and accelerating feature engineering for anomaly detection.
- Automated Python pipelines, cutting manual effort by 60% and reducing reporting time from 8 to 3 hours.
April 2023 - April 2024
Project Engineer - II / Data Scientist
WIPRO Limited | Bengaluru, India
- Executed a recommendation system that increased internal employee's engagement and skill development by 50%.
- Utilized sorting and filtering techniques to deliver personalized content effectively.
- Developed time series and machine learning models to capture and predict user preferences.
- Improved trading response times by 20% and cut risk reporting delays by 25% using Openlink Endur at Barclays UK.
September 2020 - October 2022