Simarpreet Kaur | Applied AI & ML

About

I'm a Computer Science graduate from McMaster University (Honours, May 2026) with a strong focus on applied data science and machine learning, particularly in building models and analyses that translate into real-world impact. My work emphasizes the full data science lifecycle — from data cleaning and exploratory analysis to modeling, evaluation, and clear communication of results.

I bring a detail-oriented and analytical approach, with experience working under real-world constraints such as data quality issues, validation, and model reliability. I value reproducibility, thoughtful evaluation, and practical deployment over purely theoretical results.

Alongside the technical work, I lead operations for a high-volume retail location — owning P&L review, inventory accuracy, hiring, and team development. That side of my background is where most of my project ideas come from: I see operational problems firsthand, then build systems to fix them. It also means I'm used to explaining technical decisions to people who care about outcomes rather than models.

GitHub LinkedIn

Experience

Team Lead, Operations

Best Buy · Specialist → Senior Specialist → Team Lead

Oct 2022 – Present

Review P&L and key business metrics to identify performance gaps and direct operational priorities for a high-volume location.
Investigate inventory discrepancies to root cause, driving improvements in stock accuracy and shrink control.
Lead hiring, onboarding, training, and coaching for the operations team.
Own daily operations across inventory, fulfillment, merchandising, and customer service.
Build internal tooling to replace manual operational processes — see the Inventory Tracking Tool below.

Data Science Analyst

Outlier

Oct 2025 – Present

Built data science workflows in Python for LLM benchmarking, covering EDA, feature engineering, and model evaluation.
Ran error analysis and model performance assessment (F1, ROC-AUC, cross-validation) using Pandas, NumPy, and scikit-learn.

Selected Projects

Inventory Tracking Tool

Internal

Barcode Scan • Web App • Deployed in Production

High-value stock was leaving a secured cage with no record of who took what or when. Built and deployed a browser-based scanning tool that captures UPC, employee, and timestamp at the point of removal — running on a shared store device, in daily use by staff.

Shipped around real constraints: no access to the corporate app platform, mobile-browser-only deployment, UPC integrity through CSV export, and type-ahead employee lookup after watching staff scroll a long list.

Internal operations tool — walkthrough available on request.

Pocket AI

Private

Whisper AI • NLP • Speech → JSON

Real-time wake-word detection pipeline that captures voice input with sub-500ms latency, transcribes via Whisper AI, and converts 90% of free-form speech into structured JSON logs capturing task IDs, actions, timestamps, and assignees.

Private client project — architecture and technical details available upon request.

Multilingual News Explorer

SBERT • Semantic Search • NLP • Flask API

Cross-lingual semantic retrieval and summarization app over multi-locale news feeds. Retrieval across 8 languages with 80% text reduction and sub-200ms query responses via a PyTorch-backed Flask inference API with precomputed embeddings.

View Repo

Procurement KPI Analysis

SQL • Pandas • Tableau • EDA

End-to-end EDA on a Kaggle procurement dataset using SQL and Pandas to analyze supplier performance and key operational KPIs. Tableau dashboards surface procurement trends and identify cost-saving and efficiency opportunities.

View Repo