Rishabh Shah Email & Phone Number

New York, Ny, Us

Seattle, Wa, Us

Seattle, Wa, Us

Seattle, Wa, Us

San Francisco, California, Us

• Implemented a fast DBScan clustering algorithm by using caching to precompute distances using FAISS an efficient similarity search library. This enabled analysts to faster label fraudulent content and improved client retention.• Extracted customer data using PostgreSQL. Analyzed impact of clustering on those customers to validate fraud capture.• Built a prototype tool in React, Redux & Flask to surface clustering for demo purposes to customers and the sales team.• Used the prototype to gain buy in from senior leadership on the value of doing clustering. • Assisted in GDPR deletion by preserving a GDPR compliant dataset for data scientists to use.

Seattle, Wa, Us

- Working on segmenting Caenorhabditis elegans to better predict genes that increase longevity of life using Keras and TensorFlow. - Building software that identifies death events of C. elegans based on time series images of their movements.- Generated segmented masks of worms based on annotated bounding boxes of C. elegans using image processing.

San Francisco, California, Us

• Worked on Sift’s Machine Learning systems to help fight fraud by finding spam users in the digital trust and safety space.• Built an Apache Airflow DAG to schedule a weekly job to update data bases by importing data from updated data sources. Reduced time taken to complete the same tasks from 1 day to 2 hours by using batch updates and avoiding overheads.• Extracted distance between two different entities using concepts of flow programming in Java that improved ROC-AUC curves which resulted in better classification.• Built an image pipeline for faster experimentation using images and avoided unnecessary storage using Spark and AWS.• Used image and word embeddings from ResNet and BERT along with logistic regression and Xgboost to identify fraud users based on images resulting in improved classification for platforms heavily dependent on images using PyTorch and Scikit-Learn• Gained significant domain knowledge of the different types of fraud and spam faced by a wide range of users

Mumbai, Maharashtra, In

-Worked on deep learning algorithms in Caffe for semantic segmentation on histology images to expedite the drug discovery process for the largest drug manufacturing company in India(Sun Pharma Pvt. Ltd).-Re-purposed a deep neural network architecture called DeepLab v2 in Caffe for histology images leading to better results-Wrote a loss layer in Caffe using Python where in I implemented the forward and backward pass for dice loss in 2D images in Python which helped to improve the accuracy.-Derived the equations for the backpropagation algorithm through the loss layer.-Worked on ways to mitigate class imbalance which resulted in 30% faster training time of the models.-Made a python program using multi-threading that sped up stitching of multiple tiles into a single image by 66%.

Ahmedabad, Gujarat, In

-Collaborated with two other faculty members on this project which aims to make navigation safer for the blind people.-Interviewed the visually impaired to gain an understanding of the problems faced by them.-Involved in the literature review, implementation and design of algorithms and proposal drafting.

Ahmedabad, Gujarat, In

-Conducted lab sessions where in I delivered lectures to over 60 students at a time.-Designed problems aimed at testing student's understanding of the concepts taught in class.-Kept records of lab attendance for each student in Microsoft Excel.-Stored student lab submissions in proper file format and recorded individual grades in Microsoft Excel on a weekly basis.-Assisted students with various queries over email.

Mumbai, Maharashtra, In

-Developed a deep learning based semantic segmentation approach for segmentation of products from backgrounds.-Extracted the dominant and disparate color for automatic color tagging of products using Python and OpenCV using background removal and histogram analysis.-Worked on Lab color space for automatic color tagging of products.-This module was later deployed by the company for their website.

Ahmedabad, Gujarat, In

-Assisted students in performing lab experiments.-Graded exam papers.

Ahmedabad, Gujarat, In

-Taught students planning to retake the exam for this course over the summer.-Coordinated with 2 other TA’s in organizing the curriculum and teaching different parts of the same.-Delivered lectures to over 10 students.-Over 90% of the students cleared the exam.

Master Of Science - Ms, Data Science

Bachelor Of Technology (B.Tech.), Information And Communication Technology