Umair Amir
EDUCATION
Lahore University of Management Science
B.S Computer Science
CGPA: 3.6/4.00
Relevant Coursework: Deep Learning, Machine Learning, Artificial Intelligence, Software Engineering, Advanced Programming, Data Structures, Microcontrollers,,
Operating Systems, Databases, Network Centric Computing, Algorithms
PROFESSIONAL EXPERIENCE
Software Engineer
(May 2025 – Present)
Parcel To Ship – UAE (Remote)
Built a scalable full-stack shipping marketplace with React.js, Tailwind CSS, Node.js, and SQL, enabling users to compare courier rates,
delivery times, and services.
Integrated Stripe payments, Strapi CMS, and courier APIs (Aramex, iMile) to support secure transactions, dynamic content, and real-time
logistics tracking.
Deployed the platform under a microservices-based Control Tower model, improving system reliability, scalability, and reducing deployment
downtime.
Full Stack Developer
(Dec 2024 – April 2025)
NS Autos – Dubai, UAE (Remote)
Spearheaded the digital transformation of the automotive service center by architecting and launching a comprehensive full-stack platform using
React, TypeScript, and NodeJS, transitioning the business from offline operations to a fully digital storefront for service bookings and sales.
Engineered a resilient, cloud native infrastructure on AWS, utilizing EC2 for reliable application hosting and S3 integrated with CloudFront to
optimize the delivery of high resolution service assets, ensuring low latency performance and scalability for growing traffic.
PROJECTS
AI/ML Projects:
Forest Fire Detection using Mixture of Experts | GitHub
(May 2025 – Present)
Developed a Mixture of Experts framework combining Swin Transformer + Faster R-CNN detectors for satellite, indoor, and outdoor fire
datasets. Integrated a CNN-based self-attention gating network and Weighted Box Fusion (WBF) to improve detection accuracy across diverse
environments
Alzheimer’s MRI Slice Classification | GitHub
(Mar 2025 – Present)
Built a deep learning pipeline using Swin Transformer with Enhanced SE modules for Alzheimer’s stage classification on the OASIS-2 MRI
dataset. Achieved high classification accuracy across Non-Demented, Very Mild, Mild, and Demented categories.
Edge-Aware LoRA/QLoRA Optimization (Research Work)
(Oct 2024 – Dec 2024)
Investigated flash endurance challenges in on-device LoRA/QLoRA fine-tuning and explored if system-aware update strategies (batching, delta
encoding, adaptive scheduling) can minimize write amplification.
Development Projects:
Fleet Management & Reimbursement System | GitHub
(Jan 2025 – May 2025)
Built a full-stack React Native + Node.js application for trip logging, expense estimation, and reimbursement. Integrated iMile API for
shipment tracking, Cloudinary for image storage, and AI-based trip expense prediction with role-based dashboards for admins and truckers.
COMMUNITY & LEADERSHIP
Director of IT, LUMS Entrepreneurial Society (YLES)
Directed the planning and execution of many online events at YLES (Pakistan’s largest entrepreneurs summit), leading the end-to-end
management of digital platforms and ensuring seamless tech operations for 1,000+ participants.
TECHNICAL SKILLS
Programming Languages: Python | C/C++ | SQL | JavaScript | TypeScript | HTML/CSS
Machine Learning & Deep Learning: PyTorch | TensorFlow | Keras | Scikit-learn | FastAI | OpenCV
NLP & Multimodal AI: Large Language Models (LLMs) | Retrieval-Augmented Generation (RAG) | LangChain | Prompt Engineering | Text-toSpeech (TTS) |
Computer Vision: Vision Transformers (ViTs) | Swin Transformer | GANs | CycleGANs | CNNs | Object Detection | Semantic Segmentation
Data Analysis & Visualization: NumPy | Pandas | Matplotlib | Seaborn | Power BI
Development Tools & Frameworks: FastAPI | Flask | Streamlit | React.js | Next.js | React Native | MERN | Spring Boot | Tailwind CSS |
MongoDB | PostgreSQL | Firebase | Redis | Docker | CI/CD
