Disha Punjabi
Big Data EngineerSoftware Engineer at Google
Google
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Software Engineer at Google
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
University of Southern California
Masters in Science, Computer Science
January 1, 2015 – January 1, 2017
Veermata Jijabai Technological Institute
Bachelor of Technology (B.Tech.), Computer Engineering
January 1, 2008 – January 1, 2012
Software Engineer
March 1, 2021 – Present
Roku Inc.
Sr. Software Engineer ( Machine Learning, Big Data)
July 1, 2019 – March 1, 2021
Roku Inc.
Data Science & Big Data Intern
May 1, 2017 – May 1, 2017
Roku Inc.
Data Engineer (Machine Learning)
January 1, 2017 – March 1, 2021
Roku Inc.
Big Data Intern
May 1, 2016 – August 1, 2016
Los Gatos CA
Credit Suisse
Software Developer, Algorithm Execution Services
July 1, 2012 – May 1, 2015
Mumbai Area, India
Recommendation Engine for a Question Answer Portal
September 1, 2016 – December 1, 2016
- Build a recommendation system in order to best match a question with most relevant users (to answer it) - Trained a model to optimize on probability of a user answering a question - Using features of the users along with those of the questions, developed a hybrid recommender system, combining raw collaborative filtering with user and question feature set
Lyrics Sentiment Classification using Natural Language Processing
March 1, 2016 – Present
• Prepared the data pipeline for sentiment prediction using word-to-vector technique to facilitate building learning models • Applied Support Vector Machines to text classification problem using semantics of word to vector for lyrics • Compared the performance and accuracy against other relevant models like Naive Bayes and Logistic Regression
Hotel Recommendations Classifier (Naive bayes Multinomial Classifier)
January 1, 2016 – Present
Created a review classifier to categorize hotel reviews into two categories from: [Truthful Or Deceptive ] and one of [Positive or Negative] -Designed a Naive Bayes Multinomial Classifier, trained on three fold input data and validated on one fold - Tested the classifier on a random dataset, it generated a score of 90% (F1/accuracy score) The 'Truthful' reviews are those which are fetched from the real-world web-pages , 'Deceptive' are the ones that are generated by Mechanical Turks, 'Positive' reviews correspond to the people being happy and 'Negative' reviews correspond to the unhappy customers
Full-featured Operating System
August 1, 2015 – Present
- Worked on implementing and throughly testing major functionalities of an Intel 32 Operating System -Virtual File Systems, Process Thread Lifecycle, Virtual Memory Management - Used POSIX thread library to implement threads and mutexes for synchronization implementations
Web Scraping for Weapons Dataset
August 1, 2015 – Present
Quick Demonstration: https://www.youtube.com/watch?v=yCdcXdjVcq0 - Crawled, Parsed and Visualized weapons data from the web to answer sensitive and challenging questions that the faced by the nation by associating weapon types with regions (geotopic parsing), age and time based trends in weapon sales (Solr indexing and filtering), discovering unlawful sale of explosives with the help of tools like Nutch, Tika, Solr, Banana - Working in a team of four, I optimized crawling parameters to massage the fetched data, processed specific MIME types, customized algorithms for metadata based image similarity (Charikar’s Simhash algorithm), used term frequency-inverse document frequency for content based ranking algorithms using Solr, visualized data using Banana UI and couple of D3 (Data-Driven-Document) diagrams
Storyboard
June 1, 2013 – Present
- Storyboard provided low Latency alerting mechanism for client communication - I was responsible for transforming the application to a generic design, hence allowing decoupling the business logic from the programming aspect. This made it faster and easier to add new alerts. - Extended the real-time alerting system for trade and market based notifications - Using the C# .NET framework, worked on creating UI tools for managing the alerts from the central server - Analyzed historic and real-time trade data sources to generate back-end data consolidation, contributing towards ease of new-alert addition
Improving Image Based Search Techniques using CBIR and SIFT
July 1, 2011 – Present
- Carried out an extensive literature survey to study Image Based Retrieval Techniques- Content Based Image Retrieval (CBIR) with Scale Invariant Feature Transform (SIFT) - Proposed an algorithm to combine the two algorithms. This combination resulted in more accurate image searches owing to speed factor of CBIR and scale/ rotation invariance of SIFT. - A research paper describing the approach was published in International Journal of Engineering and Technology
Cultural Fit Analysis
The candidate's project portfolio is diverse, ranging from machine learning applications (recommendation engines, sentiment analysis) to core computer science (operating systems, image processing) and data engineering (web scraping). This breadth suggests adaptability and a strong curiosity, which can be positive for cultural fit in dynamic environments. The experience at Google and Roku Inc. aligns well with a target role in Big Data Engineering, indicating a fit with fast-paced, tech-driven cultures. The candidate's experience level (14 years) and progression to a Senior Software Engineer role at Roku and Software Engineer at Google suggest a commitment to growth and a strong work ethic. The personal projects demonstrate initiative and a passion for technology beyond professional responsibilities.
Soft Skills & Operational Fit
The candidate's project descriptions indicate an ability to work on complex problems independently and in teams (e.g., 'Working in a team of four' for Web Scraping project). The experience at Credit Suisse suggests an understanding of high-stakes environments and the need for scalable, robust solutions. The description of transforming an application to a generic design for 'Storyboard' indicates problem-solving and design thinking skills. However, without specific assessment data on communication, logical reasoning, or stress handling, a definitive assessment of soft skills and operational fit is limited.