Ahmed Malik

Global B2B IT consultancy optimizing systems for finance, SaaS, and retail clients

• Eliminated 10+ hours of manual reporting weekly by replacing legacy Excel workflows with self-serve Power BI dashboards.
• Ensured 100% data accuracy in recurring reports by converting manual SQL calculations into automated, Airflow-scheduled dbt models on AWS Redshift.
• Reduced loan-demand forecast error (MAPE) by 22% against the baseline model by training Temporal Fusion Transformer models in AWS SageMaker.
• Identified high-risk loan defaults with 87% prediction accuracy (F1) by engineering an ensemble machine learning model using PyTorch (CatBoost, NN, SVM).
• Accelerated team issue resolution by integrating a Claude debugging workflow into weekly SQL and dbt code reviews.

Civic tech nonprofit leveraging open data to enhance urban road safety

• Reduced processing time for 20K+ weekly crash records from 8 hours to 10 minutes by building an AWS (S3, Glue, Lambda, Athena) ETL pipeline.
• Streamlined the identification of cycling crashes with 92% precision by fine-tuning a MiniBERT classifier on thousands of police radio transcripts using Hugging Face.
• Prioritized 12 high-risk roads for urban safety funding by applying anomaly detection in Python to historical crash patterns.

• Accelerated applicant processing by restructuring the old database to relational database architecture (3NF) for 3K+ annual profiles.
• Saved $700K in marketing costs by clustering low-conversion prospect segments using K-Medoids in Python.
• Increased email click-through rate by 17% by running statistically powered A/B tests on message variants.
• Streamlined counselor outreach for high-probability enrollments by building a logistic-regression ranking model on historical funnel data using Python.
• Eliminated ad-hoc SQL report requests by centralizing campaign performance tracking into automated Tableau dashboards and ETL pipelines.

IT consultancy modernizing US federal and commercial systems

• Reduced API response time by 43% by optimizing PostgreSQL stored procedures for multi-million-record workflows (CRUD).

Commercial and residential property developer

• Guided localized marketing strategies by collaborating with the sales team to conduct qualitative competitor analysis and market research.
• Built a pipeline of prospective buyers through direct client networking and communicating property value propositions at promotional events.