Hello, I’m Ayesha
I transform raw data into actionable insight. With a passion for deep learning, generative AI and classical machine learning, I build solutions that make a difference.
Get my CV
Data Scientist | Deep Learning, Generative AI & Analytics
I transform raw data into actionable insight. With a passion for deep learning, generative AI and classical machine learning, I build solutions that make a difference.
Get my CV
I am a data scientist with a knack for turning numbers into narratives. After earning a Master of Science in Data Science with distinction from Kingston University, I have been exploring how generative models can transform glioblastoma histopathology. My expertise spans deep learning, generative AI, classical machine learning and statistical analysis. I thrive on solving real‑world problems across healthcare, finance, aviation, public safety and workforce analytics.
Originally from Pakistan, I have made the United Kingdom my home for more than ten years. This cross‑cultural perspective, coupled with extensive travel across Europe and the Middle East, informs both my work ethic and my worldview.
Beyond the technical, I draw on over a decade of professional experience in customer service and management. This blend of skills helps me bridge the gap between sophisticated analysis and clear business impact. Whether I’m building a data warehouse or training a neural network, my goal is always the same: to deliver insights that matter.
I work fluently with Python (PyTorch, Keras, Scikit‑learn, Pandas), R, SQL, Tableau, Power BI and AWS Glue. I design robust data pipelines, train and optimise predictive models and communicate findings through clear visualisations and dashboards. My background has fostered leadership, teamwork and communication skills that complement my analytical expertise.
Developed a bidirectional virtual staining system to translate between ISH and H&E images. Thousands of image patches were extracted and models were evaluated with FID, SSIM and PSNR. The unpaired CycleGAN model achieved the most realistic output and preserved tissue morphology.
Designed a comprehensive 22‑table relational database using Oracle APEX for a global shipping company. Modelled routes, vessels, containers, cargo and crew. Implemented primary and foreign key constraints, triggers and validation rules, and wrote SQL queries for cargo tracking and crew compliance.
Built a star‑schema data warehouse to analyse flight punctuality across UK airports. Integrated tens of thousands of flight records, cleaned and normalised the data and developed Tableau dashboards to explore delays, cancellations and route performance.
Analysed more than 306 000 crime records from 2016–2022 using Python. Cleaned data, engineered temporal and spatial features, applied K‑Means clustering and produced visualisations revealing post‑COVID declines, crime hotspots and temporal patterns.
Trained and compared eight classifiers on the Scikit‑learn digits dataset. Hyper‑parameter tuning produced a support vector machine with balanced accuracy of about 98.6 percent and ROC AUC of 99.9 percent. Performance was evaluated using confusion matrices and ROC curves.
Built LSTM, GRU, CNN‑LSTM and dense models in Keras to forecast Starbucks stock prices. Used sliding‑window sequences and scaling; the CNN‑LSTM hybrid achieved a test R² of approximately 0.889. Predictions were plotted against actual prices to assess performance.
Performed ETL on the GHCN‑D dataset using AWS Glue, Athena and CloudFormation. Created Glue crawlers, cleaned and converted data to Parquet, wrote Athena queries and deployed a reusable template with appropriate IAM policies.
Implemented multiple linear regression to predict red wine quality (R² ≈ 0.32) and logistic regression to assess hip fracture risk. Improved the AUC from 0.909 to 0.926 by incorporating height and medication variables.
Explored the SUPPORT2 dataset (9 105 patients and 48 variables) to understand patient outcomes. Assessed missingness, outliers, distributions, correlations and variable cleaning, identifying strong relationships between severity scores and mortality.
Investigated salary and satisfaction data for 1 470 employees using R. Conducted correlation analysis, t‑tests, ANOVA and chi‑squared tests. Found moderate correlation between pay and satisfaction and no significant gender pay gap.
I believe that life is short; enjoy it before it is too late. There is no time to complain; time is money. Live and let live. Learning never finishes: even on your last day you should remain curious or risk becoming mentally stagnant. I strive to keep a humble, joyful outlook: seriousness in action balanced with a light heart and a ready smile.
I have had the privilege of travelling widely, exploring cultures across Switzerland, Germany, France, Belgium, Luxembourg, Dubai, Abu Dhabi and Qatar. I grew up in Pakistan and have called the United Kingdom home for more than a decade, with travel across England, Scotland, Wales (including Cardiff) and Ireland as well as visits to numerous Pakistani cities. These experiences have strengthened my appreciation for local traditions, simple living and the importance of humanity over material wealth. I believe in embracing ordinary joys rather than chasing lavish lifestyles and respect people who value kindness and empathy.
I am open to opportunities in health data, research roles, PhD programmes, business analysis and any position where a data scientist can make an impact. After a decade of professional experience outside formal studies, my return to academia reignited my passion for data science. Earning a distinction despite this gap demonstrates my commitment and ability to learn quickly. I utilise AI responsibly to support coding and analysis, always understanding the algorithms and their implications. Ultimately I want to leave a positive mark and contribute to work that benefits humanity.