Adhish Thite

ac8qkt@r.postjobfree.com +1-980-***-**** Charlotte, NC linkedin.com/in/adhish-thite/ https://adhishthite.github.io/

Machine Learning Deep Learning Data Science Computer Vision Data Visualization & Analysis Big Data Analytics

Machine Learning Engineer leveraging expertise in Deep Learning, Data Exploration and Visualization, Application and Web Development to effectively translate client requirements for insightful, data-driven business decision making. Exhibits proven ability to optimize business processes through cutting-edge analytics, winning project leadership skills, and industry expertise of frameworks such as TensorFlow, Keras, and tools like Tableau, SAS, Google Cloud Platform & Analytics Services.

Core Competencies: Deep Learning (TensorFlow, Keras, PyTorch), Machine Learning & Data Analytics (Python, R), Data Science Stack (Pandas, scikit-learn, NumPy), Cloud Computing (Hadoop, Spark), Cloud Platforms (AWS, GCP, Azure), Data Visualization (Tableau, Google Charts), Statistical Analysis (SAS, SPSS, Excel), Google Analytics, Database Management (SQL), Application Development (Java).

Professional Experience

Machine Learning Engineer (Intern), GoCollect (Charlotte, NC) 2/19 – Present

AI-based UX Enhancement: Engineered a Computer Vision-based Deep Learning pipeline autonomously on Google Cloud Platform to accurately identify a comic book based on an image, decreasing search time by 30% as compared to text-based search.

Predictive Analytics: Predicted the sales price range of comic books on a 5-figure scale by analyzing a dataset of 33 million records from the GoCollect platform in Python using TensorFlow.

Machine Learning Engineer (Intern) in Computer Vision, Welch Labs (Charlotte, NC) 10/18 – Present

Modular Algorithm Design: Partnered with Microsoft & SpyGlass to reduce the false positive identification rate of defective windshields by 100% by implementing an ensemble of Convolutional SVMs and CNNs using OpenCV and Keras in Python.

Increased Savings: Projected to save USD 1 million per quarter. Targeting to optimize the model to have an 8s response time during inference after deploying as an Azure Machine Learning Web Service (REST API).

Machine Learning Engineer (Intern), Zuora, Inc. (San Francisco, CA) 6/18 – 8/18

Business Process Improvement: Reduced Zuora’s live support agent involvement by 75% by building an NLP Topic Modelling pipeline in Java and Python to correlate customer support tickets with internal knowledge base content.

Driving Customer Engagement: Accelerated ticket response time by 90% by leveraging Deep Learning algorithms to automate access validation to Salesforce.com orgs in Zendesk.

Application Development Analyst Salesforce.com Specialist and SME, Accenture (Pune, India) 4/15 – 7/17

Application Development & Maintenance: Led the end-to-end delivery of an e-commerce platform for Splunk. Acted as the Lead Salesforce.com Developer and Technical Team Lead for the Salesforce’s first-ever implementation of a cloud-on-cloud model.

System Overhaul: Facilitated complete overhaul of a Purchase Order flow by developing key delivery components. Augmented Sales Reps efficiency by 50% by deploying advanced automation processes via Salesforce.com customization & configuration.

Education

Master of Science (Computer Science) University of North Carolina at Charlotte May 2019

Courses: Machine Learning, Visual Analytics, Big Data Analytics, Cloud Computing for Data Analysis.

Bachelor of Engineering (Computer Engineering) University of Pune, India May 2014

Courses: Algorithms, Data Structures, Operating Systems, Theory of Computation, Artificial Intelligence.

Academic Projects

Improved Decoupled Neural Interfaces [Individual Research]: Reduced the training time for Deep Neural Networks by 50% by implementing an independent ‘pre-training’ module in TensorFlow. Created a weight initializer based on input-input mapping.

Neural Image Caption Generator: Generated best-fit captions for given images by implementing a VGG-16 + LSTM model in Keras. Optimized model while securing a 5% increase in BLEU translation score by using the Inception module on reduced vocabulary size.

Spoken Digits Audio Classifier (Sound-MNIST): Correctly identified spoken digits by developing a 97% accurate Convolutional Neural Network Classifier with PyTorch. Represented audio data in a numeric format by leveraging the Mel-Frequency Cepstrum Coefficient.

Contact this candidate