Return to Portfolio

Mihir Ramdevputram

AI & Data Science Student · Software Builder

B.Sc. (Hons.) AI & Data Science undergraduate at IAR University with hands-on industrial training as a Python Developer Intern at Codec Technologies Pvt. Ltd. Creator of VisionSphere and SentiVision-AI. Passionate about computer vision, NLP sentiment pipelines, and applied software engineering.

Ahmedabad, Gujarat, India
·
beyondmihir@gmail.com
·
GitHub
·
LinkedIn

Education //

B.Sc. (Hons.) Artificial Intelligence & Data Science

2026 – 2030 (Expected)
IAR University · Gandhinagar, Gujarat, India
  • • Curriculum covering Python, Relational Databases (SQL), C programming, Calculus, and Applied Statistics.
  • • Developing applied AI and computer vision software projects alongside academic coursework.

Senior Secondary Education — Class XII (CBSE, PCB Stream)

2024 – 2025
Airport School · Ahmedabad, Gujarat, India
  • • Completed Physics, Chemistry, and Biology stream with analytical rigor (72.4%).

Secondary School Examination — Class X (CBSE)

2022 – 2023
Airport School · Ahmedabad, Gujarat, India
  • • Graduated CBSE secondary curriculum with 80.0% aggregate.

Industrial Training & Internship Experience //

Python Developer Intern

Sep 2026 – Oct 2026 (1 month)
Codec Technologies Pvt. Ltd. · India
  • • Completed 1-month AICTE & ICAC-approved industrial internship program focused on Python software development.
  • • Engineered VisionSphere: An explainable image classification platform with Grad-CAM visualization supporting 1,000 ImageNet categories.
  • • Developed SentiVision-AI: A multimodal text sentiment and facial emotion analysis application utilizing RoBERTa and OpenCV.
  • • Awarded Certificate of Completion and reported Letter of Recommendation from Program Manager Dr. Anurag Shrivastava.
PythonFlaskOpenCVTensorFlow/KerasRoBERTaGrad-CAMREST APIs

Engineering Projects & Deliverables //

VisionSphere — Explainable Image Recognition Platform

GitHub Repo ↗

A comprehensive computer vision web platform for classifying images across 1,000 categories with explainable attention heatmaps.

  • • Integrated MobileNetV3, ResNet50, and EfficientNet-B0 architectures for comparative model inference.
  • • Implemented Grad-CAM heatmaps allowing visual inspection of convolutional layer activations.
  • • Built OpenCV diagnostic pipeline with filter presets, ROI cropping, and batch classification.
  • • Exposed clean REST API endpoints and data export functionality.
PythonFlaskTensorFlowOpenCVGrad-CAMBootstrap/CSS

SentiVision-AI — Multimodal Sentiment & Emotion Analysis

GitHub Repo ↗

A multimodal AI platform combining Twitter-RoBERTa text sentiment analysis with real-time facial emotion recognition.

  • • Built an NLP ensemble combining CardiffNLP RoBERTa, VADER, and TextBlob for text analysis.
  • • Integrated OpenCV facial detection with FER emotion classification into live video streams.
  • • Developed interactive dashboards with Chart.js and PDF summary report exports.
PythonFlaskTransformersOpenCVChart.jsMongoDB

Technical Competencies //

Languages:Python, JavaScript (ES6+), SQL, C, C++, Java, HTML5, CSS3
AI & Computer Vision:TensorFlow, Keras, OpenCV, Grad-CAM, Transformers, RoBERTa, VADER
Web & Frameworks:Flask, REST APIs, Next.js, Tailwind CSS, Node.js
Databases & Tools:MySQL, MongoDB, Git, GitHub, Docker

Verified Credentials & Certifications //

Python Developer Internship Completion CertificateCodec Technologies Pvt. Ltd. (AICTE & ICAC Approved)
2026
AI Fluency: Framework & FoundationsAnthropic
2026
Ethical Hacking & Cybersecurity WorkshopVaultofCodes
2026