Vinay Patil | SDE

Hello Explorer,

If we are in simulation, then about me pages will be specification charts:
and here is mine,


- I am Vinay Patil, and my program has a lot of roles to play!. - I am a master's student at Carnegie Mellon University pursuing my dream by exploring the world of Machine Learning and Data Science. - I am also a passionate Software Developer who loves to play with frameworks and technologies to make things start working 😜. - Apart from the world of machines, I am a cyclist, a newbie photographer, and a puzzler. Let me tell you a fun fact all images from this website are captured by me :). - As the saying goes "time is money", hence I am optimizing my introduction with a specification chart below. feel free to optimize further with the search bar above.
Actively looking for jobs?	Yes, Full-time - SDE/ SWE/ ML based roles - starting May 2022
What I am studying?	masters in Electronics and Computer Engineering focusing on Machine learning
Courses at CMU	Sem-1 (GPA: 3.22): 10601 : Introduction to Machine Learning 18786 : Introduction to Deep Learning 18741 : Computer Networks Sem-2 (GPA:4.0, CGPA:3.61): 16-720 : Computer Vision 18-785 : Data Inference and applied Machine Learning 18-819 : Introduction to Quantum Computing and Machine Learning Sem-3 (current): 18-733 : Applied Cryptography 18-787 : Data Analytics 18-788 : Big Data Science 15-295 : Competition Programming and Problem Solving
Expected Graduation Date	May 2022
Resume	Resume
Undergraduate details	Bachelor's of Engineering in Computer Engineering with 8.82/10 CGPA from Mumbai University, India
Currently working	As Research Assistant at CMU : Machine Learning for PCB designs
Work Experience	2 years & 7 months as Software Engineer at JP Morgan Chase and 3 internships in software development

Things I practiced MOST

Programming Languages :-
Java, Python Advanced
Angular,JavaScript Intermediate
Kotlin, React, C Basic
Framework/ Platforms :-
Android, Firebase Advanced
Numpy, Scikit-learn, Pytorch, Pandas, IBM MQ, Kafka Intermediate
AWS cloud Basic
Tools :-
Git, Maven, Jira board Advanced
Database :-
MySQL, MongoDB, Firebase real-time database Advanced

Projects

Although github is best place to showcase projects, but because of few restriction some repos are private. So providing a brief intro to few of my projects:-

Experience: Research project

- Worked on designing efficient printed circuit board design optimized to minimize IR-drop across the board
- Implemented evolutionary algorithm combined with deep network and metaballs to improvise A* based baseline solution. Which reduced execution time by 50% and improved the convergence rate
- Collaborated in experiments by adding PCB designs and performed a comparative study of A* and new approach
- Tech specs [Python, Pytorch, Sk-learn]
- Algorithms [Genetic algorithm, Metaball designs]

Experience : JP Morgan - instruction capture project

-Collaborated with senior engineers and product owners to create smart solutions for trade instruction capture
- Developed a generic solution that will support multiple clients with minimal efforts, this reduced on-boarding and logistic time needed for new clients from few days to one day
- Upgraded core transformation engine to support advance data formats such as proto amps and Json which saved manual coding, code replication and reduced the development time from 2 sprints work effort to 3 days
- Coached three new joiners in the team, by introducing and helping out with technical details of the project
- Received Q4-2018 (Excellent performing new joiner) and Q2-2019 (Excellent performer in the team) awards within the team of instruction capture space
- Tech specs [Java, SQL, Gaia-Cloud, Linux, React, Angular, JSON, XML, Protobuf, Swift, IBM-MQ, GIT, Jira board]

Quantum Image Classifier

- Implemented image classifier based on quantum hadamard edge detection with the quantum image encoding
- Experimented on various quantum devices like Dwave, Qiskit, and simulators from amazon brackets.
- Implemented baseline models using computer vision Harris-cornerd detector combined with K-means
- Developed baseline models using Deep learning Resnet networks.
- Tested on MNIST and CIFAR dataset and achieved 77% and 45% accuracy
Tech-specs: [Python, Qiskit, AWS bracket, Resnet, Bag of words, K-means, Harris-corner detector]

Object Tracking in Videos

- Implemented Lucas-Kandae optical flow detection to detect and track selected objects from the video
- Improvised implementation by handling anomalies using a combination of Matthew-baker and Lucas-Kanade method
Tech-specs: [Numpy, PyTorch]

Image stitching

- Implemented planar homography based image stitching application with BRIEF descriptors to generate panoroma.
- Implemented poisson image stitching to blend two images and form a smooten merged image
Tech-specs: [Numpy, PyTorch]

Object detection using neural network

- Implemented deep network to detect objects within an image and assign them visual words using bag of words technique.
- Tested the implementation using nist36 achieved 95% accuracy.
Tech-specs: [Numpy, PyTorch], Datasets:[nist], Models:[MLP]

Optimizing Deep Neural Networks

- Designed Compact and Efficient Deep Neural Networks by a combination of pruning and quantization.
- Optimization achieved by pruning un-used connection by intelligent analysis of gradients & weights.
- The network is further optimized by performing k-means clustering over weights which helped reduce the memory requirement
Highlights: reduced memory usage by 10% and speed-up the model by 5%, along with mean drop in accuracy by 3-4%.
Tech-specs: [Numpy, PyTorch], Datasets:[CIFAR-10 & 100], Models:[ResNet34, ResNet50, VGG16]

Face Recognition

- Built a classification model for images with multiple classes and a verification model to find similarities between group of images using PyTorch.
- Implemented core architecture of resnet34 from scratch combined with data engineering this model achieved 90% accuracy in both tasks
Tech-specs: [Numpy, PyTorch, torchvision], AWS Ec-2.

Sentiment Polarity Analyzer

- Designed a Natural Language Processing system in NumPy, to detect the tone of the sentence, trained on movie reviews data in supervised fashion using Logistic Regression which used sparse matrix implementation which reduced the timing for training by 25% without losing the accuracy.
Tech-specs: [Numpy, PyTorch, torchvision], Google Colab.

Project for NGO

- Reorganized school systems operated under NGO with solution to track student progress and teacher’s appraisal
- Provided daily/weekly/monthly intelligent report system which help NGO to predict funding and resources logistic.
- Collaborated with six other team members to deliver an android and web solution as part of “Force For Good” program under J P Morgan chase News letter muktangan-goes-mobile/ Tech-specs: [Angular, Firebase, JSON, Android, Bootstrap].

Search Engine for private documents

- Implemented a 3-stage search engine to scan-analyze-use the document, designed based on term frequency-inverse document frequency algorithm
- Improvised the algorithm to solve its three anomalies, had published research paper for same in IRJET Journal with 7.52 impact [ref: P-ISSN: 2395-0072]
- Built a chatbot to intelligently answer the questions based on a ranking system and extracted keyword knowledge from the documents
Tech-specs: [Python rest-api, Firebase, CSS], highlights:[TF-IDF, Page ranking, NLP, PCA, context search ], Dataset:[SQL, real-time Firebase DB]

Other Noteworthy projects

Technical Farmer:- runner's up application from hackathon, end-to-end solution for NGO, consumer and Farmer for :- Android, Nodejs, hosting
RAIT CSI publicity Ticketing Application:- Android Link
Remote File manager:- Client-Server File system to store,update,delete,accessing file:- Java,socket programming server|Client
Video Streaming server: Java, Tomcat Benchmarking
Digital Document Generator:- university-wide forms auto-fill and submission system. Java8.0, JavaFX scene builder Link

Get In Touch

"Everyone you will ever meet knows something you don't" ~ Bill Nye
I really believe in this quote, so I always keep my inbox open, whether you have a question or just want to get in touch I'll try my best to get back to you!
Also, I will really appreciate if you can drop me mails about any relevant job updates.