About Me

Amey Ambade

I am a Data Scientist with experience in machine learning, computer vision and software development. I build cross-domain end-to-end AI/ML solutions for customers. I have developed innovative and actionable ML product prototypes and operationalized data science solutions by collaborating with different product lines and business teams. I'm part cynic, part hopeful. I like to write short stories, play the uke and cook healthy stuff. My record time for keeping a houseplant alive is five days.

My Career

Schlumberger

I work with international oil companies to help build machine learning workflows to improve their in-field efficiency. I am involved in analyzing interesting patterns in data, building detection and prediction models around them and deploying them using custom pipelines. Frameworks: Tensorflow, scikit-learn, Python, Keras, R, Dataiku.

2019 - present, Houston
Data Scientist

Schlumberger

I designed a deep learning-based model to detect and classify erosion in drilling and measurement TeleScope tool to aid experts on rigs automate tool monitoring. I modeled and developed an LSTM model for predicting drill cracking based on time series data from bottom hole assembly sensors. I built a health analyzer application for predicting magnet subsystem sensor failures for PowerDrive tools. Frameworks: Tensorflow, scikit-learn, Python.

2018, Houston
Data Scientist, Summer Intern

Columbia University

I'm graduated with a Master's degree in Computer Science with a specialization in Machine Learning. My goal during my degree was to gain sound technical knowledge to help me fuel my desire to learn and improve myself as computer science professional. My courses reflect my interests and along with my academic profile are indicative of my technical prowess.

2018, New York
Graduate student in Computer Science

Indian Institute of Technology (IIT)

I earned a B. Tech. in Computer Science and Engineering with a Minor degree in Humanities and Social Sciences. My focus was on building a well-rounded knowledge base of computer science fundamentals and software engineering practices to contribute toward my career in the technology industry. I was also involved in managing student extracurricular activities: I served as the Chief Editor for all publications under the university literary club and I helped found the university astronomy club.

2017, Indore
Computer Science and Engineering, Undergraduate

École Internationale des Sciences du Traitement de l'Information

I built automatic data extraction and analysis models using NLP and machine learning for AREL, an e-learning platform used by 1500+ French students and alumni.

2016, Paris
Software Engineer, Summer Intern

TeamChat by Webaroo Inc.

I integrated widely used web services like Slack, Pocket, Yandex and Wikipedia with Teamchat's SDK and API platform using Java and RESTful APIs. I was responsible for real-time notification and translation systems for direct access to data from Teamchat's web or mobile interface, that are commercially available to clientele today.

2015, Mumbai
Software Engineer, Summer Intern

My Projects

Legal Claim Prediction Software

I'm developing software to predict the outcome of contractual damage litigation to benefit academics in debates, parties in drafting contracts, counsels in their pre-trial exchanges and judges in assisting their rulings. I am using NLP to extract information from online legal databases and ML techniques to design the prediction model.

Contextual Neural Attentive Recommendation Machine

I built a deep learning based architecture to improve session-based recommendations for e-commerce scenarios where user profiles are invisible using an attention mechanism and context information about the user's actions. Achieved significant improvements in Recall@20 and MRR@20 scores over the present baseline methods on the YouChoose dataset.

Pictionary using PictoGANs

I designed and implemented a Generative Adversarial Network model called PictoGAN that can play Pictionary. I modeled multi-class data from the Quick, Draw! dataset to allow the generator network to produce output similar to human drawings of different objects.

Integro

I developed a Django web app to pull information from various services like YouTube, Wikipedia, Twitter and Reddit based on searches, and make recommendations using AWS DynamoDB and Lambda. I designed content-based and collaborative filtering-based recommendation systems and AWS Cognito for credential management and trained the model on Apache Spark with the Alternating Least Squares algorithm.

ameyxd.github.io

Thiss portfolio website built with Jekyll and GitHub Pages. It uses the beautiful Jalpc theme.

...

AI Customer Service App

I developed a Customer Service chatbot, an iOS application that included facial verification, and allowed users to have complete conversations with a bot for booking a cab, a hotel or a parking spot. I built the service using AWS frameworks like Rekognition, Lex, Cognito, API Gateway, Lambda, Kinesis, DynamoDB, S3 and Elastic Beanstalk.

Object Recognition using CNNs

I led the team working to develop an artificially intelligent system that can detect objects posing as a security threat in environments under surveillance. I designed a Python-based supervised learning algorithm on a CUDA-Theano-Keras backbone.

Modeling 3D rooms

I built a 3D model of a home with interactive objects and developed game tactics based on movement and placement of those objects using OpenGL and C++.

Analysis of face detection techniques

An implementation of the Viola-Jones face detection technique and analysis of its shortcomings in MATLAB. Used Haar classifiers and eigenfaces for training to detect presence, location and orientation of faces in images.

My Blog

Link

Ambade Amey

Data Scientist | Machine Learning