GAGANDEEP SINGH

Well-qualified Data Scientist experienced working with vast data sets to break down information, gather relevant points and solve advanced business problems. Skilled in Key Information Extraction from highly Unstructured Documents with NLP and OD techniques. Offering 3.5 years of experience in improving business operations.

• Have developed Invoice Document Parsing Model with 95.5 meF1 score utilizing state of the art technique.
• Worked in extensive research for key Information extraction from all type of unstructured documents.
• Developed chat-bot system for bill of lading and letter of credit documents.
• Prediction of Viscosity from PENETRATION value and reactor profile using TensorFlow platform ( Neural Network custom architecture). Model achieved MAE of 150 poise. against the 2400 to 3600 poise range for VG30 Grade Bitumen.
• Using Time Series Forecasting using LSTMs, Prediction of Intermediate Properties (Flow of Reduced Crude Oil) 30 minutes early, so that downstream plant can be adjusted without any disturbance. Model correctly predicts RCO flow within 2% of total flow.
• Stabilization of Hydrogen grid Pressure using Time Series Forecasting.
• Worked on Data Extraction From Unstructured Documents.
• Currently working on Anomaly Detection model for a Critical Turbine in CCR ( Catalytic Cracking Unit), to improve performance and prior prediction of maintenance.

• Two time campus winner for HUL Techspark. An Inter-IIT technical competition with case study Based on Environment friendly implementable solution with working model demonstration
• Winner of the event 'Green Tech' during Keramos'16 (Department of Ceramic Engineering). Solution presented for Plastic minimisation with green alternatives
• Winner of the event 'Rachna' during Udbhav'16 (Department of Chemical Engineering). An process optimisation Event
• Awarded the 'Rajeev Gandhi Digital Student Scholarship' for meritorious students in Senior Secondary Examination

Piano, Cricket, Racing sport and Photography