Mushtaq Patel

MUSHTAQ PATEL ML Engineer/Data Scientist https://www.linkedin.com/in/mushtaq-patel-/ - https://github.com/mushtaqpatel0505 -,- https://medium.com/@mushtaqpatel Indore, India PROFESSIONAL SUMMARY Experienced Machine Learning Engineer with proven success in building successful algorithms and predictive models for different industries. Adept at end to end pipeline design & development including data gathering, data cleaning and preprocessing, data analysis and visualization, model training, model deployment, and testing. Handle various projects in different domains. A passionate and thriving engineer with the ability to apply ML techniques and algorithm development to solve real-world industry problems and provide cutting-edge solutions. Love to work on edge intelligence. AREA OF EXPERTISE ✅ Classification and Regression ✅ Clustering ✅ Natural Language Processing ✅ Computer Vision ✅ Machine Learning Algorithms ✅ Deep Learning ✅ Image Classification, Segmentation and Detection✅ Data Analysis and Visualization ✅ Data Mining ✅Web scraping ✅ Time Series Analysis ✅ Recommendation System ✅ Statistical Analysis ✅ Predictive Analysis TECHNICAL SKILLS Tools • Python, R • C/C++ • Java, JavaScript • Hadoop, Spark, SQL • Linux, REST • Jupyter Notebook • MySQL Packages • SciKit-Learn • NumPy, SciPy, Pandas • Matplotlib, Seaborn, Plotly • NLTK, Beautiful Soup, Statsmodel • Keras, TensorFlow, PyTorch • OpenCV, Tesseract • Flask Statistics/Machine Learning • Linear/Logistic Regression, SVM • Naïve Bayes, KNN, Decision Tree • Ensemble Models, XgBoost, GBDT • DNN, CNN, LSTM/RNN, GAN • Transformers, Auto-encoder • KMeans, DBSCAN, Hierarchical • YOLO, SSD, ARIMA, Prophet SOFT SKILLS • Critical Thinking, Problem Solving • Adaptability • Eager to learn • Communication • Storytelling • Team Player • Curiosity • Business understanding PROFESSIONAL EXPERIENCE Zenith Engineer Inc Lead Data Scientist Pune,IN Sep’ 21 to Dec’ 21 ● Real estate property price prediction for USA beach areas Created the machine learning model tom predict the price of land near by sea beaches in USA Grubbrr Systems Pvt Ltd Machine Learning Engineer Ahmedabad,IN Jun’ 21 to Aug’ 21 ● Food Item detection, counting and tracking at checkout Trained SSD mobilenet from scratch to detect the food item, count them and track them at checkout to get the total price using the camera of the store. ● Food item forecasting for restaurants Forecasting the different food items should be sold on next day(Morning, afternoon, Evening, night) using the previous historical and weather data CERTAINTY INFOTECH PVT LTD Machine Learning Engineer Indore,IN Sep’ 20 to May’ 21 ● Web Page Optimization and Prediction Used Reinforcement Learning’s Thompson sampling to optimize the web pages and wrote an algorithm to predict the best page to show to the customer to increase the sales. ● Built dashboards and metrics on grow.com Built the dashboards and metrics on grow.com using data analysis skills. Created datasets using SQL. ● Information extraction from bank cheques Used pytesseract and trained the CNN-LSTM model on Handwritten english words. Extracted the Information from cheques like Name, Amount etc. F(X) DATA LABS PVT LTD Machine Learning Engineer Ahmedabad,IN Jan’ 20 to Aug’ 20 ● Smart-Kitchen Used the Tensorflow object detection model to detect the chefs in the kitchen who are violating the rules of the kitchen. Also, send emails of rule violations with photos to the corresponding manager in real-time. ● Fraud detection in replacement of parts of machines Labeled the data using K-Prototype clustering and trained ensemble models to detect frauds in replacement of parts. And also predicted failure dates of parts in the future using Neural Networks and deployed the model using Flask and Restful API. ● Business-oriented solutions to increase the revenue of a salon Used data analysis to find the trends or patterns in customers’ behaviour, predicted the number of customers on each day, and each month by using time series forecasting and managing the staff according to that and used K-Means Clustering for customer segmentation into three categories to increase customers and revenue of salon. ● Sentiment predictions with the reasoning of food reviews of mixed languages Predicted rating of mixed Indian languages food reviews using KerasClassifier and also gave reasons for predictions using LIME. Used pre-trained BERT for word embeddings. ● Image dimensionality reduction for different images comparison in low latency system Used Auto-Encoder to decrease the dimension of images up to 75% without losing much information and with high accuracy for comparison with other images. ● Data-Driven Decisions to Improve Ranking on the online platform Used various data science approaches to help a client improve his rank on this platform. Performed web scraping for data extraction with Beautiful Soup and used matplotlib and seaborn for Data Analysis and Visualization. Used Linear regression for prediction of optimal fee for services and rank on webpage. AAIC PVT LTD Data Scientist Hyderabad, IN July’ 17 - July 18 ● Apparel Recommendation System (NLP and Computer Vision) Recommend similar apparel products in e-commerce using product descriptions and Images. Classification of images of clothing into 10 different classes done using Hyper-parameter tuned Dense and Convolutional Neural Networks. ● Redefining Cancer Treatment with Machine Learning (NLP) Uses Logistic Regression, Naïve Bayes, and SVM to automatically classify genetic variations of cancers using text reports. ● Human Activity Recognition using smartphone data (Sensors data) Build a Neural Network that predicts human activities such as Walking, Walking-Upstairs, WalkingDownstairs, Sitting, Standing, or Laying using gyroscope and accelerometer readings. ● NVIDIA Self Driving Car Paper Implementation (Computer Vision) Uses Convolutional Networks to predict steering angle according to the road. Here features are images (middle, left, right), speed, throttle, and brake. Implemented model using TensorFlow and Keras ● Emotion based songs recommendation Used facial expressions data like happy, sad, angry, etc. and Haar-cascade classifier to detect face and trained CNN classification model and recommended songs according to expression. Also decreases the size of the model so that it can be deployed on low energy devices. EDUCATION B.E. (Computer Science and Engineering ) Ujjain Engineering College Aug’ 12 - Jun’ 17 Ujjain, IN ● GPA 6.4/10 10+2 (Physics, Mathematics, Chemistry) 2010 ● 80% 10th Higher Secondary ● 87% 2008 CERTIFICATION Coursera ● Neural Networks and Deep Learning ● Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization Applied AI Course ● Applied AI/ Machine Learning ACHIEVEMENTS ● Secured 35 th rank in “Machine Learning for IoT” competition out of 7250 individuals and 50 teams. ● Secured 111th rank in “Mobile Analytics” competition out of 6345 individuals and 63 teams. TECH-STACK EXPERIENCE ● Python/R/SQL ● Machine Learning/Deep Learning ● TensorFlow/PyTorch/Keras ● NLP/OCR/Computer Vision/Time-Series ● Data Analysis/Visualization/Modeling 6 years 4 years 4 years 4 years 4 year