William Scott Paka
Data Science Enthusiat | X-Fullstack Developer
- Email: williamscott701@gmail.com
- Mobile: +91-7845108300
Successful professional with 2 years of work experience as "Head of Development" in a start-up, leading a highly skilled team of 8+ employees and interns in building a robust Spring MVC e-commerce application. Currently a Masters student in IIITD specializing in Artificial Intelligence. Enthusiast in Data Science, with demonstrated abilites in AI / ML / DL / IR / NLP / CV having developed solid real-world projects. Published a paper on “Multitask Learning for Blackmarket Tweet Detection” . Current research works include developing a Novel Topic Based Generative Model to find the sentiment of text given a topic. Successful in intense and demanding environments, providing decisive team leadership with a track record of being highly motivated and goal oriented.
Python
Java / PHP / ASP
Ionic / Angular
C++
J2EE / Spring MVC
PyTorch / Keras / Scikit-Learn / NLTK / SWN
GitHub / Bluemix / AWS
Applied AI with Deep Learning – IBM
Under: Dr. Tanmoy Chakraborty, IIITD
In this project we implemented Multitask Learning model using cross stitch units to detect tweets that are posted in blackmarket. We used primary task to be detection of blackmarket tweet and the second task is to predict the likes and retweets of the tweet. We published this work in Asonam, 2019.
Under: Dr. Saket Anand, IIITD
In this research project we explored ways to identify individuals in Monkeys and Macaques which is considered a difficult problem statement. We have implemented few of the solutions in which we can learn a pair wise loss using KL Divergence. We have performed the tasks on classification, open set and closed set and compared the results.
Under: Dr. Tanmoy Chakraborty, IIITD
In this project we aim at suggesting the user the relevant answer from stack overflow by using textual and code similarities from query. We made the first attempt to make use of the adjacency matrix of the codes’ Control Flow Graph to find multiple similarity measures that can be used together to find a similarity measure.
Under: Dr. Saket Anand, IIITD
In this project we explored few Deep Learning techniques to detect and classify abusive comments. Dataset: Wikipedia detox, Twitter. Achieving F1 ~ 85%.
Under: Dr. Mayank Vatsa, IIITD
In this project we implemented and compared multiple AI models which can play the game 2048 to make the maximum score. AI Models: Search Algorithms, RL, NN and NN+RL. Our search algorithms even exceeded the Stanford paper results.
Under: Dr. R. Elijah Blessing, Director Collaborations, Karunya University
Stipend: Rs 20,000 – 30,000 / month (1 dollar per hour paid monthly) Worked as, a ‘Software Engineer’ on a part-time basis for a consultancy project on ’Java web development’ for a US based company.
1. YOLO PyTorch – Object Detection and Classification (Real Time)
2. SOTA EfficientNets (Transfer Learning)
2. AutoEncoders for Faces
2. LSTM for sequential audio data classification
3. Neural Networks from Scratch – (MNIST Dataset)
4. Linear Regression from Scratch – (Boston Housing Dataset)
5. Logistic Regression from Scratch – (MNIST Dataset)
6. SVM Classification – Hindi Handwritten Digits
7. Alexnet in PyTorch – CIFAR Dataset
Worked as ‘Head of Development’ in our start-up Covenant IT Solutions Private Limited efficiently leading a team of 8+ employees and interns, trying to build a robust and stable online e-commerce with IBM Bluemix Cloud and many other technologies.
Worked as Web Designer and Developer at Computer Technology Centre, Karunya University under “Earn While You Learn Scheme” for a fixed stipend.
In this research project we aim to develop a novel generative model architecture which can find the sentiment of a statement given a text and a topic.
This project is to Design a guest access solution to provide network access to visitors segregating the traffic from the internal network using web-based authentication with full auditing of location, MAC, IP address, username.
1. Ranked 14, in Hard level Hacker Rank Machine Learning Challenge – Guess the Flipkart Query
2. 2nd prize in COD++ (2015), a coding Event at a National Level Technical Symposium ‘Expertise 2k15’ Conducted by Department of IT, Karunya University
3. Received grade 10/10 in all programming labs (Java Programming Lab, Cryptography and Network Security Lab, Web Technology Lab and, Digital Design Lab)and 9/10 in all the other major labs.
Authorised Publisher for TDS - Towards Data Science, a premium of Medium for Data Science field in AI, ML, DL, IR domains. (February 2019 – Till Date)
Team Member of IIITD Lab – Data Science (LCS2) working on Novel Approaches to “Detect Collusive Tweets in Twitter”. (December 2018 – Till Date)
Additional Director and Authorised Signatory for start-up “Covenant IT Solutions Pvt Ltd”. (January 2018 – Till Date)
Head of Development for start-up “Covenant IT Solutions Pvt Ltd” efficiently leading a team of 8+ employees and interns. (March 2016 – October 2018)
Senior Residential Advisor in Boys Hostel, Karunya University. (March 2016 – March 2017)
Student Coordinator for Photoshop event named 'Picster' for a National Level Technical Symposium in Karunya University. (March 2015)
Website Admin for TECHAGON’15, a National Level Technical Symposium organized by Department of CSE, Karunya University. (March 2015)
Website Admin for MAD (Making A Difference) Event by Dove Department, Karunya University. (Auguest 2015)
Photography
Reading, Reading, Reading
Graphic Designing