Dev Vaghasiya
Machine Learning Engineer & Data Scientist
Who I Am
A passionate Machine Learning Engineer and Data Scientist with expertise in building production-grade ML systems, generative AI models, and scalable web applications. Currently pursuing a B.S. in Data Science and Application at IIT Madras, with a strong foundation in Electronic & Communication Engineering from GEC Surat. My experience spans from developing diffusion-based image generation models and deploying GPU-accelerated ML workloads on Kubernetes, to building real-time web applications with Next.js and React. I'm committed to leveraging cutting-edge technologies to solve real-world problems and deliver impactful solutions.
What I Do
Languages
5 technologies
Web Development
5 technologies
ML & GenAI
7 technologies
MLOps & Deployment
5 technologies
Cloud & Infrastructure
3 technologies
Databases
3 technologies
Tools & Analytics
5 essential tools & platforms
Featured Work
Quick Chat
Developed a real-time chat application with authentication and group/private messaging capabilities
Tech Stack
Outcomes
Supports 100+ concurrent users with ~50–100 messages per second throughput
Home Finder
Developed a real-time home and PG finder application with intelligent roommate matching
Tech Stack
Outcomes
Supports 200+ property listings with a fully responsive and intuitive user interface
Facial Recognition Student Attendance System
Built an automated face-recognition attendance system using computer vision and machine learning
Tech Stack
Outcomes
Optimized preprocessing and dataset pipelines to improve recognition accuracy by ~15–20%
Flipkart Grid 6.0 Robotics Round-2
Built and trained a CNN-based image classification model for identifying fruits, vegetables, and perishable items
Tech Stack
Outcomes
Achieved ~85–90% accuracy on a custom dataset of perishable items
Where I've Worked
Machine Learning Intern
Beyond Karma Tech
January 2026 - Present
Surat, Gujarat
- •Developed diffusion-based image generation models using Hugging Face and analyzed 2.2K+ social media ad records to identify high-conversion audience segments, improving category-specific targeting by ~15–20%
- •Built and deployed production-grade ML pipelines using Docker with CI/CD automation, reducing manual deployment effort by ~40% and improving release reliability across model iterations
- •Deployed GPU-accelerated ML and LLM inference workloads using Kubernetes, implementing vLLM-based serving to achieve ~30% lower inference latency on AMD MI300X GPUs
Data & AI Skills Intern
Microsoft Power by CloudThat
July 2, 2025 - July 17, 2025
Surat, Gujarat
- •Completed Azure AI Fundamentals (AI-900) certification and built NLP solutions with Azure AI Language
- •Developed generative artificial intelligence (AI) models in Azure Machine Learning
- •Performed data analysis using Microsoft Power BI Data Analysis (PL-300)
My Journey
B.S. Data Science and Application
Indian Institute of Technology, Madras
B.E. Electronic & Communication Engineering
GEC, Surat
Achievements
Microsoft Power BI Data Analysis (PL-300)
Microsoft
Azure AI Fundamentals (AI-900)
Microsoft
Tata Imagination Challenge 2024 Student Track
Tata
NIT Surat Data Science Workshop
NIT Surat
Let's Connect
vaghasiyadev84@gmail.com
Phone
+916353628633
Location
Surat, Gujarat
LeetCode
devvaghasiya
linkedin.com/in/dev-vaghasiya
GitHub
github.com/Vaghasiya-Dev