Michael Bamisi

Built an end-to-end regression pipeline to forecast product sales across DSN Mart retail locations, supporting inventory optimization and supply chain planning. Engineered features including price-to-category ratios, price per kg, item visibility ratios and within-store price percentiles, plus frequency encoding and categorical interactions. Handled missing values and zero-value anomalies using group-based imputation and missingness indicators.

Data Science Nigeria (DSN) AI Bootcamp Qualification Hackathon (2026-10)

• Built an end-to-end regression pipeline to forecast product sales across DSN Mart retail locations, supporting inventory optimization and supply chain planning.
• Engineered features including price-to-category ratios, price per kg, item visibility ratios and within-store price percentiles, plus frequency encoding and categorical interactions.
• Handled missing values and zero-value anomalies using group-based imputation and missingness indicators.
• Trained CatBoost/LightGBM models with 5-fold cross-validation and early stopping, and ensembled the fold models for final predictions.

(2026-09)

• Built an end-to-end classification pipeline to detect fraudulent credit card transactions on a highly imbalanced dataset.
• Applied SMOTE to rebalance the classes and Z-score outlier removal to clean extreme values before modeling.
• Trained and developed ( Logistic Regression, Random Forest, and XGBoost models) evaluating performance using Precision, Recall, and F1-score to optimize for imbalanced class distributions rather than accuracy.

Quant Analyst & ML Engineer (2026-06)

• Developed an algorithmic framework to identify Break of Structure (BOS) and Fair Value Gap (FVG) in financial markets.
• Implemented automated back-testing methodologies wing Python to evaluate the effectiveness of Smart Money Concepts (SMC).
• Accelerated Step Search Script
• Engineered & Python script (accelerated_step_search) to optimize data retrieval and processing speeds.

Database Design Engineer (2026-03)

• Designed a clean, 9-table PostgreSQL schema (3NF) to structure municipal zones, user roles, and report life cycles without data duplication.
• Optimized high-frequency analytical queries and transactional endpoints by implementing 7 custom B-tree indexes across key foreign keys and timestamp ranges (idx_tickets_reporter, idx_tickets_status, idx_tickets_created, idx_locations_wardzone).
• Engineered multi-table database transactions with parameterised SQL to ensure full ACID compliance and automatic rollback handling across simultaneous inserts for tickets, GPS location records, and audit status logs.
• Built an immutable audit logging architecture (status_log) to track status changes, actor IDs, and timestamps, preventing data tampering and providing full traceability for municipal service operations.

Photography and Video Content Volunteer at Build54 (2026-08)

Captured high-quality event photography and produced dynamic video content to document key initiatives, workshops, and community engagements.

Welfare and Hospitality Volunteer at Fernwood (AWS) (2025-06)

Managed guest welfare, hospitality, and on-site logistics to ensure smooth event execution I also executed promotional and outreach strategies to drive event awareness, resulting in increased attendee turnout.

Bachelor of Science in Mathematics and Computer Science – University of Lagos (2021)