Kshitij Patil is a senior undergraduate student studying Computer Engineering at Pune Institute of Computer Technology. He has experience working on medical imaging and deep learning projects. His areas of focus include computer vision, deep learning frameworks like TensorFlow and PyTorch, programming languages like Python and Java, and deploying machine learning models on mobile and web platforms. He has participated in several hackathons and coding competitions, winning some of them.

1.
Kshitij Patil Senior Undergraduate,
Dept. of Computer Engineering,
Pune Institute of Computer Technology
kshitijpatil98@gmail.com
+919423706080
www.linkedin.com/in/kshitijpatil98
www.github.com/Kshitij09
https://kshitij09.github.io/dev-blog/
EDUCATION
B.E. Computer Engineering
Pune Institute of Computer Technology
Aug 2017- Present
CGPA: 8.3
Diploma in Information Technology
Government Polytechnic, Kolhapur
Aug 2014 - May 2017
Percentage: 90.88%
EXPERIENCE
Periwinkle Technologies, ​Pune - ​Project
Intern (Medical Imaging)
Sept 2019 - June 2020
- Develop a visual evaluation algorithm for
Screening Cervical Precancer/Cancer using deep
learning-based multimodal systems.
PICT ACM Student Chapter, ​Pune ​—
Technical core committee member &
Domain Director of Android
Sep 2018 - Sept 2019
- Guiding fellow PASC members by conducting
seminars and mentoring projects. Keeping them
abreast of recent trends in software
development.
CoReCo technologies, ​Pune — ​Intern
May 2016 - June 2016
- ​Explored Docker technology with the
deliberate study of containerization. Ascertained
how the effective usage of Docker containers
could reinforce application deployment.
PROJECTS
Handwritten text recognition system
(June’20 - present)
- Working on a cursive handwritten text recognition
system. Currently evaluating the feasibility of several
APIs and have attained the initial results of fine-tuning
tesseract OCR on the IAM dataset.
- ​Technologies Used: tesseract-ocr, PyTorch
Cervical Cancer Screening
(Sept’19 - present)
- A visual evaluation system for cervical cancer
screening based on the VIA image and demographic
data. Given the limited amount of data and overlapping
classes, we explored several pre-training techniques to
learn discriminative features. The best model was
further evaluated using Explainability algorithms to
support its decision.
- ​Technologies Used: Pytorch, fastai2
CNN based Forest Fire Detection
(Nov 2019 - Mar’20)
- A CNN based Forest Fire Detection algorithm intended
to be deployed on camera-enabled edge devices. Dataset
was created by extracting frames from YouTube videos
and by aggregating several resources. ​[​source code​]
- ​Technologies Used: Tensorflow, Pytorch, fastai, OpenCV
Crowd Counting for disaster management
(Aug 2018 - Sep 2018)
- Developed a system based on the novel deep learning
architecture - CSRNet (Y.Li et.al. CVPR '18) to output a
crowd density map corresponding to an input image,
and hence deduce the count from the density map.
- ​The model was then deployed on an Android platform
using a quantized version of it (TFLite).
- ​Technologies Used: Tensorflow, Keras, Android, Firebase

2. SKILLS
Deep Learning: Computer Vision, a thorough
understanding of standard deep learning
architectures, activations, optimizers, and loss
functions.
Deep Learning frameworks: ​TensorFlow, Keras,
PyTorch, fastai
Programming: ​Python, Java, Kotlin, C++, Swift
Front end development: ​Angular
Backend development: ​Spring-boot, Flask
Database: ​MongoDB, Mysql
Miscellaneous: ​Android, Firebase, Docker
- Able to deploy machine learning models on
Android, web, and edge devices using the
TensorFlow ecosystem.
- Design and deploy Microservices using Docker
RELEVANT COURSES AND CERTIFICATIONS
CS20 -​Tensorflow for Deep Learning
Research
Stanford University
Machine Learning - Stanford
University
Coursera
Grade achieved: 95.7%
Convolutional Neural Networks -
deeplearning.ai
Coursera
Grade achieved: 98.4%
Google Cloud Training​ -
Qwiklabs
1. GCP Essentials
2. Baseline: Data, ML, AI
3. OK Google: Build Interactive Apps with
Google Assistant
CS231 - Convolutional Neural Networks
for Visual Recognition
Stanford university
Assignment solutions (2019) - [​source​]
Practical Deep Learning for Coders​:
part-1 & 2
Fast.ai
ACHIEVEMENTS
Winner: MindSpark Hackathon 3.0:
College of Engineering, Pune
- Part of a 4 member team who stood first
amongst 52 participant teams at Mindspark
Hackathon, the biggest technical event in Pune.
Winner: Software Development - Credenz’18:
PICT IEEE Student Branch, Pune
- ​Winner of senior category in software
development at Credenz, organized by PICT IEEE
Student Branch(R10) for the headcount
monitoring system.
Software Development - Credenz’17:
PICT IEEE Student Branch, Pune
- ​Winner of junior category in software
development at Credenz, organized by PICT IEEE
Student Branch(R10) for the project ‘FrameIT’.
CodeTrix - Avishkar 2016:
Government College of Engineering, Karad
- Winner of the coding competition organized by
the Government College of Engineering, Karad.
Best Outgoing Student: (2017 batch)
Government Polytechnic, Kolhapur
- Recognized as ​‘Best Outgoing Student’​ of IT
Department from Government Polytechnic,
Kolhapur
EXTRA-CURRICULAR ACTIVITIES
Speaker at Punecommunity.AnitaB.org
- Delivered a session on ​‘Exploring the World of
Android Apps’​.
ORGANIZATIONS
fastai
- Active member on fastai forums, answers deep
learning and fastai2 related questions. ​[​profile​]