Sagar Kumar
Software Developer & Machine Learning Engineer.
The Architecture
of My Skills.
2
Tech Internships3+
Leadership Roles100+
Members MentoredB.Tech CSE Undergrad at AUGN
I aim to build a holistic understanding of the technology landscape by working across enterprise systems, product development, and application development. From SAP-based solutions to building and enhancing AI/ML models and applications, I focus on exploring and evolving new technologies with a strong systems-level perspective.
Core Stack
- Python
- Dart
- C++
- Java
- C#
- SQL
- SAP ABAP
AI & ML Ecosystem
- TensorFlow
- PyTorch
- Keras
- Computer Vision
- OCR Pipelines
- Google Gemini API
- OpenAI API
- Claude API
- NumPy & Pandas
- Scikit-learn
- Neural Network Architecture
Frameworks & Tooling
- Flutter
- Firebase
- MoviePy (Video Automation)
- Tkinter (Desktop GUI)
- Enterprise Cloud Architecture
- Next.js
- Brevo SMTP
Experience &
Leadership.
Machine Learning Engineer Intern
AICTE (IBM SkillsBuild)
- Developed and evaluated supervised and unsupervised learning models in Python.
- Applied core ML techniques through project-based implementations and mentor-guided technical evaluations.
SAP ABAP Developer Intern
CorporateServe Solutions Pvt. Ltd.
- Assisted in ABAP development and debugging across SAP modules.
- Supported code reviews, performance tuning, and documentation.
Student Research President
AUGN Research Society (IETE)
- Leading a 60+ member research-focused student body organizing AI/ML workshops and technical discussions.
Technical President & Core Member
Fusion Tech Club
- Mentoring 100+ members across AI/ML initiatives and technical competitions.
- Organized key events: Ideathon 2.0, Codathon 3.0, and E-Hackathon.
Vice President & Co-Founder
ASET E-Cell
- Co-founded the Entrepreneurship Cell to support student-led startups and innovation.
- Organized multiple hackathons and entrepreneurship workshops.
Selected
Innovations.
Reinforced Sparse Attention (RSA)
Designing a reinforcement-driven sparse attention mechanism to reduce Transformer self-attention complexity from O(n²) to near O(n) for long-sequence modeling. Evaluating sparsity policies, computational efficiency, and performance trade-offs across large-scale tasks.
Transformer vs LSTM Review
Built and evaluated LSTM and Transformer models for sentiment classification. Conducted a comparative study analyzing accuracy, training efficiency, parameter complexity, and generalization performance.
VIDGEN YouTube Automation
A complete YouTube automation tool designed for educational content creation. Leverages Google Gemini to automatically generate programming Q&A, and MoviePy to dynamically render short videos with custom overlays. Includes secure API key rotation and automated public uploading processes.
E-Waste ImgClassifier
An application designed to classify electronic waste by assessing device damage. Detects broken devices, classifies damage severity (intact, minor, moderate, severe), and estimates an overall damage percentage. Delivered via a simple web interface and integration-ready API endpoints.
Neural Network from Scratch
Implemented a fully connected neural network in NumPy with absolute manual backpropagation. Compared SGD, Momentum, RMSProp, and Adam on the MNIST dataset, analyzing convergence and generalization behavior.
App & Cloud
Ecosystems.
ANGEN Desktop AI
OCR Assistant
A cross-platform desktop application that lets users select any screen region to extract text via OCR, instantly sending it to Gemini, OpenAI, or Claude APIs for real-time answers directly in-app.
Rentastic
Local Marketplace
A cross-platform mobile ecosystem dubbed "Amazon for rentals", enabling users to seamlessly browse, post, and manage peer-to-peer rented items with secure authentication and advanced filtering on iOS and Android.
Vaapiskaro
SaaS CRM Ecosystem
A multi-tenant startup CRM platform designed to streamline corporate workflows. Built with Next.js and Firebase, it features dynamic role-based dashboards tailored for Admins, Sub-admins, and Customers, integrating permission controls, client management, and automated communications powered by Brevo SMTP.
Education &
Credentials.
Amity University
B.Tech in Computer Science & Engineering
Greater Noida • CGPA: 7.82 / 10 • (2023 - 2027)
Coursework: Linear Algebra, Probability & Statistics, Machine Learning, Design & Analysis of Algorithms.
DAV Public School
Ghaziabad • (2016 - 2023)
Licenses & Certifications
Let's Build.
Currently exploring new ML engineering and software development opportunities.