Shubham Baid

AI and Computer Vision Engineer

4+

Years Experience

5+

Certifications

2

Publications
Get In Touch LinkedIn Profile Download Resume

About Me

Hey! I'm an AI and Computer Vision Engineer, driven by a passion for harnessing the limitless potential of AI to create impactful change. I specialize in developing and optimizing cutting-edge computer vision pipelines and integrating Vision Language Models (VLMs) & Large Language Models (LLMs).

My core expertise lies in C/C++, Python, and advanced libraries like OpenCV, PyTorch, TensorFlow, and TensorRT. I thrive on building robust, efficient AI solutions and staying ahead of the curve in this rapidly evolving field.

My work is backed by certifications including Nvidia Jetson AI Specialist, Intel Edge AI Specialist, and TensorFlow Developer, reflecting my commitment to continuous learning and practical application.

Computer Vision Deep Learning Python C/C++ PyTorch TensorFlow TensorRT OpenCV LLMs / VLMs Generative AI Edge AI Model Optimization Rapid Prototyping
Shubham Baid Profile

Work Experience

AI Engineer
CAFU
January 2025 - April 2025 (4 months)
Engineered agentic workflow marketing content generation powered by LLMs, delivering 15% CTR increase and 50% reduction in copywriting workflow time through automated content optimization.
Optimised machine learning-based ETA prediction system with real-time analytics, decreasing delayed fulfillment from 13% to 8% (3-month period) and reducing breaches by 42% for delays >5 minutes.
Designed automated B2B lead acquisition with autonomous agents and multi-modal LLMs, processing 1,400 prospects monthly with minimal human intervention. Generated 4 high-conversion leads while achieving 90% reduction in manual prospecting hours and substantial CAC improvement.
Senior AI Engineer
Avathon
May 2022 - January 2025 (2 years 9 months)
Engaged in integrating Multimodal Language Models (LLMs) to optimize traditional CV models accuracy.
Led the development of VAIA's School Safety Suite, implemented use cases including firearm detection and person re-identification across multiple cameras which is now deployed across hundreds of cameras across multiple schools in the US.
Leading a team of AI Engineers and guided them to adapt right practices for streamlined development and timely deliverables.
Designed a resource-efficient computer vision pipeline for low-power ARM hardware, optimizing for limited CPU (50%) and RAM utilising NEON and other libraries, incorporating hardware-level enhancements and model quantization.
Led the development of a gamified auto-image annotation tool, enhancing dataset preparation and model training efficiency through active learning concept.
Senior Artificial Intelligence Engineer
Integration Wizards Solutions
December 2021 - May 2022 (6 months)
Developed an ALPR (Automatic License Plate Recognition) module with high acceleration and minimal resource usage footprint, compatible with both high-speed CPU and GPU inference.
Worked on onboarding various computer vision models to an accelerated inference pipeline for GPUs, resulting in faster and more accurate results.
Contributed to the acquisition by Sparkcognition, a leading AI and machine learning company.
Artificial Intelligence Engineer
Integration Wizards Solutions
June 2021 - December 2021 (7 months)
Modified neural network architectures to optimize for specific use-cases, improving accuracy and performance.
Trained custom computer vision models at scale for warehouses, resulting in increased efficiency in HSE.
Artificial Intelligence Intern
Integration Wizards Solutions
September 2020 - May 2021 (9 months)
Trained backbone models for PPE detection, vehicle body-type classification, and fall-arrester detection.
Deployed computer vision solutions at scale, demonstrating a practical approach to solving complex technical challenges.

Publications

Text Generation Tool for Writing Assistance using Transformer.

JAZ 2023, Vol 44, p938 April 1, 2023

Writing assistance tool build using transformer. Submitted for publication in August 2020, apparently got approved in 2023.

NLP Transformers Generative AI
See Publication

Detection of Different Degrees of Skin Burn using YOLOv3

TEST Engineering and Management May 1, 2020

Computer Vision YOLOv3 Medical AI
See Publication

Projects & Publications

YetiCoach

Developed a tool providing skiers with data on ski angles and gap at different steepness levels using only an action camera.

Computer Vision Sports Tech Hackathon

Aegir - Ocean Debris AI

Platform using CV & DL on satellite/UAV data to locate, classify, and predict ocean debris trajectory, aiding cleanup efforts.

Computer Vision Deep Learning Edge AI NASA Space Apps

Argus

Autonomous drone system using DL & CV to generate critical data (locality density, population estimates, depth data) for natural disaster areas.

Deep Learning Computer Vision UAV Disaster Tech

Hailey

AI-based writing assistant leveraging Hugging Face Transformers and GPT-2 for text generation and sentence completion.

NLP Transformers GPT-2 Generative AI

AutoSnotBot

Automated drone using custom YOLO model to detect whales and collect snot samples for algal bloom prediction.

Object Detection YOLO UAV Environmental AI

Cap for Blind

Arduino-based cap with ultrasonic sensors providing haptic feedback to assist visually impaired individuals.

Arduino IoT Assistive Tech Hardware

Text Generation Tool for Writing Assistance

Developed a transformer-based text generation tool to assist with writing tasks, improving content creation efficiency.

NLP Transformers Publication

Detection of Different Degrees of Skin Burn using YOLOv3

Implemented YOLOv3 architecture to accurately detect and classify different degrees of skin burns, aiding medical diagnosis.

YOLO Medical AI Publication

VAIA's School Safety Suite

Led development of comprehensive safety monitoring system with firearm detection and person re-identification across multiple cameras.

Object Detection Re-ID Production

Automatic License Plate Recognition

Developed high-performance ALPR module with minimal resource footprint for both CPU and GPU inference.

OCR Optimization Production

Education & Certifications

Bachelor of Technology (BTech), Computer Science
REVA University
2017 - 2021

Professional Certifications

Intel Edge AI Certification

Certification focused on optimizing and deploying AI solutions on Intel hardware for edge computing.

Nvidia Jetson AI Specialist

Specialized certification in deploying AI models on edge devices using Nvidia Jetson platform

TensorFlow Developer Certificate

Professional certification in building and training neural networks using TensorFlow

Convolutional Neural Networks

Advanced training in CNN architectures and computer vision applications (Coursera)

Neural Networks and Deep Learning

Comprehensive certification in neural network fundamentals and deep learning techniques (Coursera)

Awards & Recognition

Track Winner - Hack Zurich 2022

Awarded for winning a specific track at Europe's largest hackathon.

Track Winner - NASA Space Apps 2022

Recognized for winning a track in the international NASA hackathon.

Best Outgoing Student - 2021

Recognized for academic excellence and contributions to REVA University

Hackfest 2019

2nd Prize in National Level Hackathon conducted by REVA University

Sentinal Hack 2.0

2nd position at State Level Hackathon conducted by KSIT

Galactic Problem Solver

Two-time recipient (2019, 2020) for completing NASA Space Apps challenges.

Latest Blog Post

Make your Deep Learning models infer at Minato speed in Python.

Published on: Sun, 02 May 2021

Read More
View All Posts

Get In Touch

✉️

Email

shubhambaid99@gmail.com

LinkedIn

linkedin.com/in/shubhambaid21

🤝

Let's Collaborate

Open to discussing AI projects and opportunities