A versatile computer engineer with a strong background in Machine and Deep Learning, Data Analytics/ Visualization and Big Data. 3+ years’ experience as a data engineer, 2+ years of experience as a director of machine learning programs. Self-starter with a strong sense of ownership, keen to learn and ability to hit the ground running rapidly. Skilled in data cleaning, statistical modeling, real-world problem solving and programming.
AREAS OF EXPERTISE
Machine Learning: Regression, Support Vector Machines, Decision Trees, K-nearest Neighbors, Clustering Analysis, Dimensionality Reduction, Neural Networks (Deep Learning), Time series Analysis, NLP, ARIMA, Gradient Boosted Trees, PCA – Factor Analysis, Clustering, SVM, CNN.
Programming Languages: Python, iPython Notebook, R, SQL, C
Libraries: Scikit-learn, TensorFlow, Keras, Plotly, PySpark.
Databases: PostgreSQL, MySQL
Tools: MATLAB, LabVIEW
BI tools: Tableau, Adv. Excel
Big Data Analytics: AWS
Development Boards: Arduino UNO, Raspberry Pi
DIRECTOR OF MACHINE LEARNING PROGRAMS, HADE TECHNOLOGIES, ORLANDO, FL MAY 2017-PRESENT
HadePlatform.com develops investment research products, services, and APIs with artificial intelligence and its own exclusive big database with MatriX.HadePlatform.com
•Currently developing credit card fraud detection system for HADEPAY’s payment processor using K-means, logistic regression and neural networks.
•Currently building a new product, VERIFIE, a verification system to check the genuineness cryptocurrency wallets.
•Managed a team of 4 data analysts and engineers to create and deploy models for quarterly and long-term revenue outlooks. These models were used to bolster financial projections for investment experts
•Built predictive model using multivariate regression model using prodigious financial data. Achieved ~60% better results than experienced wall street analysts.
•Created 3+ machine learning models for risk assessment for stocks and crypto investments to help clients in building their portfolio.
•Built a sentimental analyzer to assess the influence of investment experts and google search influence vs crypto price. The analyzer predicted crypto price 14% more accurate than normal regression algorithms.
GRADUATE STUDENT ASSISTANT, UNIVERSITY OF CENTRAL FLORIDA., ORLANDO, FL JAN 2017-AUG 2017
•As Team Leader at Center of Research in Computer Vision team, led an image annotation project, training datasets for raw images to create captcha detection system. Increased image captcha detection accuracy from ~81% to ~92%.
DATA ANALYST, CYBERNET.IT, PUNE, MH MAY 2013- MAY2016
•Created a data scrapping APIs for business intelligence products which paved way for ‘early to market’ plan for those products. Increased profit margins by ~3% per month.
•Created data visualizations using excel, tableau, matplotlib etc. for marketing and business development team to assess data and create meaningful business insights.
•Built HTML dashboards for data insights, data derivations for future references, which yielded ~9% more sales in 11 weeks.
•Created databases through MySQL for software/hardware products and services; making a smooth transition from bulky excel sheets to organized data warehouse.
•Implemented linear and logistic regression (statistical) models for client consulting, which proved helpful across all teams to forecast and visualize future plans.
MS COMPUTER ENGINEERING - UNIVERSITY OF CENTRAL FLORIDA, ORLANDO, FL 2016-2018
-SVM Modeled Car Crash Avoidance System (Tools: MATLAB, Python 3, Support Vector Machines)
-Credit Card Fraud Detection System (Tools: Python3, Excel, Neural Networks and Logistic regression lib)
-AMES House Price Prediction Project (Tools: Python3, Excel, Linear, Bayesian and Ridge regression lib)
-“E- Commerce” recommendation system such as Netflix recommendations ‘Because you watched…’ (Tools: Python3, Excel, TensorFlow)
-Fault Line-Phase-Type Detection System (Tools: MATALB, Python3, Excel, Logistic Regression)
-State Estimation (Electrical Engg.)
BE (INSTRUMENTATION & CONTROL) - UNIVERSITY OF PUNE, PUNE, INDIA, 2011-2015
INFANT BABY MONITOTING SYSTEMS AND DATA ANALYTICS (Final Year Project Team of 2)
Hardware design: Arduino UNO, Temperature sensor, CO2 sensor, Heartbeat sensor
Software design: Linear and Logistic regression algorithms pre-trained with Real World Data.
The lethal parameter setting improved by 7% when tested with virtual environment.
Cheapest infant monitoring system built, 1 Innovation of the year award.