Software, data, and research work built with care

ML Experiment Control Center

Lightweight internal ML platform for launching experiments, tracking run metadata, comparing runs, browsing artifacts, and exporting summaries.

• Designed around reproducibility with run IDs, config versioning, seeds, and structured logs.
• Connected backend APIs, persistence, frontend dashboards, and experiment artifacts in one product-style workflow.
• Shows stronger software engineering and product thinking than a notebook-only ML project.

Model Export, Verification, and Benchmarking Toolkit

Python package and CLI for exporting models, checking output parity across runtimes, benchmarking inference, and generating reports.

• Built deployment-minded tooling with typed interfaces, YAML batch configs, failure logs, and report generation.
• Validates numerical parity across export formats before results are trusted downstream.
• Demonstrates stronger systems engineering than a standard training-only repository.

Customer Churn and Revenue Risk Analytics

Churn modeling workflow combining classification, SQL-style analysis, feature engineering, segment-level revenue risk, and retention-oriented decision support.

• Built a reproducible pipeline for churn prediction and at-risk customer analysis.
• Compared multiple models and connected outputs to revenue-risk prioritization.
• Organized results into dashboard-ready tables and presentation-ready summaries.

Demand Forecasting and Inventory Decision Support

Store-product forecasting project focused on time-series feature engineering, model comparison, and turning forecast error into inventory planning signals.

• Built lag, rolling, and calendar-driven features for short-horizon demand prediction.
• Compared forecasting approaches and summarized reliability by horizon.
• Reported store-level and product-level inventory exposure for decision support.

Boise State University, LPiNS Lab

• Built reproducible ML workflows for computer vision and efficient AI experiments across multiple datasets and model families.
• Developed evaluation scripts, export tooling, and analysis pipelines to make results easier to validate and compare.
• Worked on quantization, edge deployment, model verification, and healthcare AI problems.
• Supported labs, debugging, grading, and technical workshops for computing and systems courses.

Playense

• Coordinated software tasks, clarified requirements, and kept teams aligned on deadlines and deliverables.
• Communicated status across technical and non-technical stakeholders.
• Helped keep project execution organized through concise documentation and follow-through.

Boise State University

PhD Student in Computing · 2023 - Present

Focus areas include computer vision, data science, efficient AI, and hardware-aware machine learning. Currently conducting research in the LPiNS Lab under Dr. Omiya Hassan.

American International University-Bangladesh

BSc in Computer Science and Engineering · 2017 - 2021

Built my foundation in software engineering, algorithms, systems, and practical implementation work across academic and project settings.

Software engineering

• Python, Java, C++, SQL, Bash
• Git and GitHub, Linux, notebooks, scripts, reproducible workflows
• Debugging, implementation quality, documentation, project organization

Data science

• pandas, numpy, matplotlib, feature engineering, exploratory analysis
• Classification, forecasting, error analysis, dashboard-ready reporting
• Decision-support workflows and business-facing analytics

ML and research

• PyTorch, computer vision, model evaluation, healthcare AI
• Quantization-aware training, efficient inference, edge deployment
• LLM evaluation workflows and experiment comparison