About Me
Researcher & engineer exploring intelligence through code — curious mind, practical builder.
Curiosity is my fuel
I’m Kaleem Ahmad, an AI/ML researcher and software engineer working at the intersection of computer vision, generative AI, and multimodal systems. My journey blends research with applied engineering — from medical imaging to intelligent platforms.
What drives me is simple: solving complex problems with clarity and impact. Whether I’m building scalable pipelines, prototyping new ideas, or mentoring students, I aim to turn cutting-edge research into solutions that matter.
Innovation Focus
Building AI solutions in computer vision, generative models, and multimodal systems — always with an eye on real-world impact.
Passion Projects
From medical imaging pipelines to open-source AI research — I thrive on projects that push boundaries and create lasting value.
Continuous Learning
Always exploring the latest in deep learning, multimodal AI, and data science — expanding knowledge to fuel better research and stronger systems.
Experience
Applying AI, data science, and engineering to deliver measurable results
Demolytics Ltd
Data Scientist
Key Achievements
- Trained transformer-based GANs and VAEs for synthetic data generation on 20M+ records, achieving 97% statistical accuracy while ensuring GDPR compliance.
- Led Xero API integration to build a fully synthetic demo organization with realistic transactions, improving client demos, training, and system testing.
Technologies Used
University of Bradford, School of Management
Teaching Assistant
Key Achievements
- Delivered online labs and tutorials in Python, data visualization, statistical modeling, and machine learning for postgraduate courses.
- Mentored diverse students through project discussions, demonstrating practical applications of scikit-learn, TensorFlow, and PyTorch.
- Reviewed and improved course materials for accuracy and clarity, ensuring better student outcomes.
Technologies Used
Data Dive Ltd
Associate Data Analyst
Key Achievements
- Built Python and Apache Spark pipelines to clean and process diverse datasets, improving downstream ML model accuracy by over 90%.
- Developed sales forecasting and anomaly detection models, enabling better business decision-making.
- Created interactive dashboards in Streamlit, Dash, Power BI, and Tableau to support data-driven decision-making for non-technical users.
Technologies Used
GoLive Consulting Ltd | BritAnalytics
Junior AI Engineer
Key Achievements
- Developed ML models predicting sales trends and customer behavior with 89% accuracy, improving SME decision-making.
- Built visualization dashboards and automated ETL workflows, cutting manual data cleanup by 45% and boosting reporting efficiency.
- Worked with cross-functional teams to provide actionable insights from predictive analytics.
Technologies Used
Projects
Research and applied projects spanning computer vision, generative AI, and medical imaging
SIAS: SegFormer-based Intracranial Artery Segmentation
Transformer-based segmentation of intracranial arteries from CT angiography using SegFormer-B0, achieving robust performance with limited medical data.
Key Achievements
- Fine-tuned SegFormer-B0 for binary artery segmentation.
- Achieved 70% mIoU on CT angiography dataset.
- Built custom preprocessing pipeline for medical images.
Tech Stack
Prompt-Driven Image Analysis
Multimodal AI system combining vision and NLP for natural language-driven image segmentation, detection, and inpainting.
Key Achievements
- Integrated GroundingDINO, SAM, Stable Diffusion, and LLaVA.
- Published preprint: arXiv:2509.08489.
- Featured by Valoris Research as a high-potential venture.
Tech Stack
Real-Time 3D Facial Reconstruction
CNN–PCA pipeline for reconstructing 3D faces from single 2D images, optimized for lightweight inference and visualization.
Key Achievements
- Developed CNN-based feature extraction with PCA compression.
- Used AFLW2000-3D dataset and 3DDFA framework.
- Enabled real-time streaming with WebGL Studio.
Tech Stack
Automated Nystagmus Detection from Video
Machine learning and signal-processing pipeline for detecting involuntary eye oscillations in video data.
Key Achievements
- Developed real-time eye-tracking pipeline for oscillation analysis.
- Classified subtypes (pendular, jerk, gaze-evoked).
- Built explainable framework for clinical screening.
Tech Stack
Early Detection of Eye Disease via Scleral Redness
Segmentation and analysis of scleral redness using SAM to extract interpretable features for early detection of ocular conditions.
Key Achievements
- Used SAM for sclera segmentation in clinical images.
- Defined interpretable redness metrics: RAP and MRL.
- Provided reproducible pipeline for research applications.
Tech Stack
Dynamic Tutoring Platform for E-Learning
Interactive platform supporting multimedia-rich authoring for online learning content.
Key Achievements
- Developed modular in-browser editor with real-time collaboration.
- Integrated text, images, audio, video, and PDFs.
- Implemented autosave and sandbox architecture for educators.
Tech Stack
Technical Skills
Expertise in AI, data science, and scalable software systems
Programming Languages
AI/ML & Frameworks
Cloud, Tools & Platforms
Data & Analytics
Achievements
Recognition through certifications and continuous professional development
Selected Certifications
Machine Learning Specialization
Stanford University (DeepLearning.AI)
ML pipelines covering regression, classification, clustering, recommender systems, and reinforcement learning. Focused on model selection, hyperparameter tuning, and reproducibility.
View CertificateGenerative AI for Data Scientists
IBM | Coursera
Applied generative models across text, image, and audio. Projects included synthetic data generation, prompt engineering, and workflow integration.
View CertificateData Analyst Associate
DataCamp
36-hour Python + SQL program with case study. Covered data wrangling, exploratory analysis, and reproducible workflows. Includes certification exam.
View CertificateCybersecurity Specialization
University of Maryland | Coursera
Hands-on labs in cryptography, threat modeling, and secure system design—applied to safe ML pipelines and sensitive data handling.
View CertificateAdditional Trainings & MOOCs
IT Support Professional
Google | Coursera
Jan 2022
Foundational IT training covering networking, OS, troubleshooting, and system admin basics.
Certificate in IT (Web & Android Development)
NAVTTC (Govt. of Pakistan)
Mar 2021
Web and Android application development training with government accreditation.
Machine Learning Specialization
University of Washington | Coursera
Sep 2020
Covered supervised, unsupervised learning and ML best practices.
Introduction to Programming & IoT Specialization
University of California | Coursera
Aug 2020
Fundamentals of programming and IoT systems development.
Get In Touch
Open to collaborations, research discussions, and new opportunities
Contact Information
Location
Peshawar, Pakistan
Resume
Download my latest resume to explore my research, projects, and experience.
Download Resume