Amit Kamthane Email & Phone Number

Pune, Mh, In

India

Centrifuge Shear Prediction | Predictive Maintenance of Failures | Potash MiningIntroduction:• Shear pin is a sacrificial piece which saves rest of the equipment. • The birds typically shear 1-2 times/day in worst performance scenario.• A shear leads to 20 mins of downtime. • Also, to avoid shearing the torque is maintained at lower than optimum range thus further limiting throughput.Problem Statement:- Develop the classification model to predict shearsProject Outcomes:• Projected EBIDTA uplift: 330k USD Cost Impact Areas:• Less downtime: <20 mins• Reduced maintenance cost: 100k• Throughput increase: 5k Tons/years

Pune, Maharashtra, India

Pune, Maharashtra, India

Pune Area, India

Current Role (Designation): Data Scientist – AI/ML Developer Roles and Responsibilities: - Build AI/ML Models- Interacting with team from different technology to maintain workflow be adhered to requirements.- Interacting with Stakeholders evaluating hypothesis and business rule decisions for feature engineering and experimental design - Taking the ownership in closing calls.- Mentoring and managing a Data Science Team, with performance evaluation and maintaining p project life cycle. Project 1: Propensity Churn Model (NBFC Client Fin-tech Domain)Build a machine learning model to assign a score to each of the leads which can be used by the company to target potential leads. A higher score would mean that the lead is hot, i.e. is most likely to convert whereas a lower score would mean that the lead is warm and will mostly not get converted. Finally, Propensity modelling determined who to target and how, which help in reducing expenses. Steps performed while building model and validating model were as follows. - Parametric Tests to verify the weightage of features- Feature Selection, cleaning data set, imputing missing values- Designing pipelines to retain data integrity- Creating Ensemble Models to handle unbalanced class problem using Stacking, bootstrap aggregation, gradient boosting etc for prediction with high accuracy. Project 2: Detecting Contamination in Edible Oils using Machine Learning Models - Made use of Fatty Acid domain knowledge for EDA to explore the trends and patterns in the data, feature engineering along with signal processing to get the best information.- Contamination analysis performed on Liquid products using statistical and AI/ML techniques.- Performed R&D on Supervised and unsupervised technique to evolve more business rules for predicting contamination.- Created mathematical models for Calibrating spectrum devices.

Pune Area, India

Bombay

Education record

Education record