👋Hello, I'm
Shubham Deshmukh
|
Building intelligent, scalable solutions.
About Me
A results-driven engineer with a passion for building intelligent systems and scalable, data-centric applications.
I am a Master's student in Computer Science at Virginia Tech with a specialization in AI and Data Analytics. My background combines hands-on experience in full-stack development, data engineering, and applied machine learning. I am driven by the challenge of solving complex problems and have a proven track record of designing, developing, and deploying robust software solutions and end-to-end ML pipelines.
From developing deep learning models for scientific research to engineering scalable enterprise applications, I thrive at the intersection of data and software. I am passionate about leveraging my versatile skill set to build innovative, efficient, and impactful products. I am actively seeking roles in Software Engineering, Machine Learning, and Data Engineering.
M.S. in Computer Science
Specializing in AI & Data Analytics at Virginia Tech.
Industry Experience
2+ years building scalable software and ML systems.
Versatile Skill Set
Expertise in Full-Stack, Data Engineering, and MLOps.
Innovative Problem-Solver
Passionate about creating data-driven, impactful solutions.
Technical Toolkit
A collection of the key technologies I leverage for software, data, and machine learning engineering.
Languages
Machine Learning & Deep Learning
Web Development & Backend
Cloud, Big Data & DevOps
Databases & Visualization
Education
My academic background and qualifications.
Virginia Tech
Specialization: AI and Data Analytics
Vishwakarma Institute of Technology
Honors: AI and Data Analytics
Work Experience
My professional journey in software development, data science, and research.
AI Software Engineer Intern
J. Craig Venter Institute - Rockville, MD
Jul 2025 – Present
- Developing machine learning models to identify COVID-19 biomarkers for enhanced diagnostic capabilities and treatment optimization.
- Implementing NCBI utilities and RESTful APIs to fetch research papers from PubMed, extracting full-text content and processing through Groq Llama model in batches for automated literature analysis.
- Creating intelligent batch processing pipelines that analyze protein sequences and mutations from open-access papers, reducing manual curation time by 80% for researchers.
Graduate Research Assistant
Virginia Tech - A3 Lab - Commonwealth Cyber Initiative - Arlington, VA
Aug 2024 – May 2025
- Developed and compared 5 Deep Learning & Vision Transformer models using PyTorch, achieving best R² = 0.89 for optically active water quality prediction.
- Curated a novel pipeline for 500k+ RGB images with astral filtering and segmentation, reducing data by 40%.
- Enhanced U-Net water segmentation using GMM validation and CUDA, reducing processing time by 30%.
Software Developer Intern
J. Craig Venter Institute - Rockville, MD
Jun 2024 – Aug 2024
- Implemented D3.js and React to enhance HSP 3.0 Data Visualization, improving researchers engagement by 40%.
- Optimized Data pipeline using Python, R, Docker, and AWS (Lambda, S3), reducing manual intervention by 70%.
- Enhanced API queries, boosting database efficiency and decreasing data processing time by 20%.
Software Developer
Wipro Ltd. - Pune, India
May 2022 – Jul 2023
- Led development of 14 enterprise applications using ASP.NET Core, ensuring 100% on-time delivery.
- Administered Tableau Server, optimizing analytics data workflows and reducing report generation time by 15%.
- Resolved 42 Retail POS issues, boosting uptime by 20% and refining transaction reliability.
Data Science Intern
IFM Engineering Pvt. Ltd. - Pune, India
Aug 2021 – Jan 2022
- Developed a Real-Time Patient Monitoring System using YOLOv4-tiny and Mediapipe, attaining 95% accuracy.
- Integrated Deep Learning models & a Flask Web App, mitigating operational issues by 30%.
- Presented research at IEEE MysuruCon 2022, demonstrating 15% faster doctor response via optimized alerts.
Featured Projects
A showcase of my work across different domains, from machine learning to full-stack development.
Career Compass
Built an intelligent career guidance platform using RAG system with LLaMA-3-70B, ChromaDB, and LangChain over O*NET Data, reducing query latency by 35%. Developed a responsive Next.js/Tailwind frontend with FastAPI-Groq backend integration, boosting user engagement by 30%. Deployed on AWS EC2 with Terraform infrastructure and Tableau analytics dashboards, improving insights access and system scalability by 25%. The platform provides personalized career recommendations based on user skills, interests, and market trends.
Human Salivary Proteome (HSP) 3.0 Visualization
Developed comprehensive interactive data visualization tools for biotech research using React and D3.js, significantly improving user experience and accessibility of the HSP dataset. Created dynamic charts, protein interaction networks, and advanced filtering capabilities that enable researchers to explore complex biological data more effectively. Implemented AWS cloud infrastructure with Docker containerization for scalable deployment.
Patient Monitoring System Web Application
Led a team to develop a real-time prediction model for patient monitoring using YOLOv4-tiny, Mediapipe, and XGBoost. Achieved 95% accuracy and improved response time for healthcare professionals. The system processes live video streams to detect patient movements, vital signs, and emergency situations, automatically alerting medical staff when intervention is needed.
Pose Estimation & Analysis System
Developed an advanced pose estimation system using MediaPipe and OpenCV for real-time human pose analysis. The system achieved 32 citations in academic research and has been used in multiple healthcare and sports analytics applications. Features include multi-person pose tracking, gesture recognition, and biomechanical analysis.
Blind Tourist Guide using VLMs
Developed an assistive system for visually impaired tourists using Vision-Language Models (VLMs) to provide real-time scene understanding and navigation assistance.
Document Summarizer
Built an AI-powered document summarization tool leveraging transformer models to generate concise and accurate summaries for long-form documents.
Want to see more?
I'm always working on new projects and exploring innovative solutions. Let's discuss how we can collaborate on your next big idea.
Let's Work TogetherPublications & Articles
My peer-reviewed publications and select technical articles.
Poseanalyser: A Survey on Human Pose Estimation
- Survey of state-of-the-art architectures for Human Pose Estimation.
- Comparison of CNNs, OpenPose, and MediaPipe on COCO and MPII datasets.
SANIP: Shopping Assistant and Navigation for the Visually Impaired
Patient Monitoring System
- Developed a system for real-time patient monitoring using YOLOv4-tiny and Mediapipe.
- Achieved 98.82% accuracy for posture classification (Sleeping, Sitting, Walking, Standing).
Sign Language Detection
Classification of ASL Alphabets and Numbers Using ORB and FAST
- Compared ASL classification methods using OpenCV feature descriptors.
- ORB with KNN classifier achieved 97.08% accuracy.
Secure Network-on-Chip Architectures for MPSoC: Overview and Challenges
An overview of security challenges in Network-on-Chip (NoC) architectures for MPSoC, including hardware Trojans, DoS attacks, and secure routing techniques.
Certifications
Recognized certifications that validate my skills and expertise.
Google Data Analytics ProfessionalCertificate
Fundamentals of Accelerated Computing with CUDA C/C++
NVIDIA
Generative Adversarial Networks (GANs) Specialization
Deeplearning.AI
Let's Connect
Ready to bring your ideas to life? Let's discuss how we can work together to create something amazing.
Send me a message
Get in touch
I'm always open to discussing new opportunities, interesting projects, or just having a chat about technology and innovation.
I'm currently accepting new projects and collaborations. Let's discuss how we can work together!