Hi, I'm Shubham Sharma.
A
Self-driven, quick starter, passionate programmer with a curious mind who enjoys solving a complex and challenging real-world problems. Expert in webrtc based technology and cloud computing.
About
I am a graduate student in Computer Science and Engineering at the Australian National University. My passion lies in problem-solving and coding and I am committed to delivering 100% effort in every project I undertake. My technical expertise spans multiple programming languages and technologies, including Java, Python, Rust and C++. I have also worked extensively with WebRTC for real-time communication applications and have experience in cloud gaming technologies. Through various internships, I have gained valuable professional experience, particularly in video streaming and data science, which has further honed my skills. I am dedicated to developing sophisticated applications that address real-world challenges and positively impact millions of users.
- Languages: Java, Python, JavaScript, C, C++, Rust, HTML/CSS, Bash
- Softwares: Git, Github, Heroku, WSL2, Docker, PowerBI, Confluence, Jira
- Libraries & Frameworks: Gstreamer, WebRTC, React, Flask, d3.js, AWS, OpenCV
Seeking an opportunity to work in a challenging position that combines my skills in Software Engineering and Machine Learning. I aim to find a role that offers professional development, engaging experiences, and opportunities for personal growth.
Experience
- Implemented and automate data flows and ETL processes for financial data, covering both YTD (Year-to-Date) and LTD (Life-to-Date) periods, leveraging TM1 and ES Financials.
- Developed BI reports and dashboards, increasing stakeholder decision-making efficiency by 30%, providing clear and actionable insights into key financial metrics.
- Built a BI solution to track and analyze Journal Entries, automating alerts with Power Automate, resulting in a 20% faster response time to increasing or decreasing entry trends.
- Tools: Power BI, Power Automate, Sharepoint, SQL, Enterprise Systems
- Developed and integrated a plugin for the server pipeline, enabling dynamic adjustments to streaming parameters such as bitrate and frames per second, resulting in a 20% improvement in streaming quality.
- Efficient packet transmission from server to client and reducing the frame and packet loss in the process.
- Enhanced a cloud gaming platform, achieving a 30% reduction in latency through their proprietary technology.
- Tools: Rust, Python, Gstreamer, C
- Successfully automated data collection pipelines for the College of Health and Medicine department, increasing data acquisition efficiency by 40% and accuracy by 30%.
- Utilized Python, along with libraries such as Beautiful Soup, to develop scraping scripts and data processing tools, reducing manual data gathering time by 50%.
- Collaborated closely with cross-functional teams to identify data requirements and implement solutions, achieving a 95% success rate in meeting project objectives.
- Tools: Python, PowerBI, Excel
- In image colorization, a subset of the COCO dataset is used which includes 15000 images.
- The strategy used involves image-to-image translation using Conditional Adversarial Networks which utilizes two loss functions: L1 loss and GAN loss. This method helps in solving the problem in an unsupervised manner (by assigning numerical values to the outputs, indicating how “real” they look).
- Tools: Colab, Python, Pytorch, Fastai
Guide: Dr. Sarabjot Singh Anand (Co-Founder, Tatras Data)
Project: Proactive Policing
- Accidental data reported from various police stations in Punjab province, India, was provided.
- The first task involved constructing an interactive dashboard to visualize incidents over time intervals. D3.js framework facilitated this. Additionally, hypothesis testing was conducted using chi-square and t-tests to derive insights from the data.
- The second task aimed to optimize traffic light signals by developing an algorithm to estimate vehicle density information.
- Tools: Python, d3.js, OpenCV, PyTorch, openstreetmaps, Javascript
Projects
Neural Image Caption Generation with Visual Attention
- Used CNN for encoding and LSTM for decoding to generate captions from images, leveraging the MS-COCO dataset. Applied beam search to improve the accuracy and relevance of the generated captions.
Skills
Languages and Databases
Python
PostgreSQL
HTML5
CSS3
JavaScript
MySQL
Rust
Java
C++
Libraries
Gstreamer
AWS
React
OpenCV
WebRTC
d3.js
chart.js
Frameworks
Flask
Bootstrap
PyTorch
Softwares
Git
Heroku
Docker
Jira
Github
PowerBI
Education
The Australian National University, COllege of Computing
Canberra, ACT, Asutralia
Study: Master of Computing
- Computer Vision
- Structured Programming (JAVA)
- Logic
- Professional Practice
- Software Construction
- Relational Database
Relevant Courseworks:
Maharaja Surajmal Institute of Technology, GGSIPU
New Delhi, India
Degree: Bachelors in Computer Science and Engineering
- Artificial Intelligence
- Algorithms Design and Analysis
- Software Engineering
- Operating Systems
- Web Development
- Computer Architecture
Relevant Courseworks:
Kendriya Vidyalaya, Pragati Vihar
New Delhi, India
Study: Central Board of Secondary Education (Class XII)
Percentage: 95% (best of 4)
- Mathematics
- Computer Science
- Physics
- Chemistry
- English
Relevant Courseworks: