I’m a Data Scientist / Machine Learning Engineer with 4+ years of experience building scalable ML systems, supporting supply chain strategy, time-series pipelines for anomaly detection, applied AI solutions for driving multimillion dollar impact.

• 🧩 Currently exploring Computer Vision, Gen AI, developing a custom CNN layer that learns my photo-editing patterns and applies similar transformations autonomously; a fusion of style transfer and personalized learning.
• 🚀 At Stanley Black & Decker, I automated procurement pipelines and A/B-tested email campaigns, driving $18M+ in measurable impact.
• ⚙️ At ADP, I engineered a dynamic thresholding system in Splunk that reduced false alerts by 85%.
• 📊 Passionate about experiments, causal inference, building interpretable ML systems. Advocate for ethical AI use in business and human context.

• Languages: Python, C/C++, SQL, R
• Frameworks: PyTorch, TensorFlow, scikit-learn, LangChain, OpenCV
• ML Focus: CNNs, Time-Series Forecasting, NLP, Reinforcement Learning, Bayesian Inference, GenAI
• Data Engineering: dbt, Autosys, Snowflake, Spark, Hadoop, ETL Pipelines
• Visualization: Power BI, Matplotlib, Plotly
• Cloud & Infra: AWS (Lambda, S3, RDS, Redshift), Azure
• Specialties: Model Optimization, Statistical Analysis, Causal Experiments, MLOps, A/B Testing

To design a custom Convolutional Neural Network layer capable of learning and replicating my personal image-editing style.
The system aims to capture transformation intent, not just color shifts; by modeling latent adjustments from before/after edit pairs.

• Data Generation: Using personal edited/unedited image pairs as supervised training examples.
• Model Architecture: CNN backbone (ResNet-style) with a custom delta layer that learns adjustment residuals.
• Loss Functions: Combination of MSE (pixel difference) and perceptual loss (VGG features).
• Training Setup: PyTorch, fastai, custom dataloader, augmentation pipelines.
• Planned Extensions: Integrating reinforcement learning to refine edit confidence, and LLM-based tagging for semantic edit context.

An “auto-edit” model capable of generalizing my editing signature; creating a personalized, generative photo assistant.

Reduced 85% alert noise and improved DevOps signal clarity using Kalman Filters and Exponential Smoothing within Splunk’s ML Toolkit. Designed a reusable ML pipeline and benchmarked ITS experiments for model reliability.

Automated data aggregation from SAP ECC forecasts, Snowflake, to accelerate supplier negotiation reports by 90%. Delivered A/B experiments and improved profitability by $12M+.

• 🛰️ Customer Churn Prediction: Built an end-to-end ML pipeline using XGBoost, FastAPI, Docker, and MLflow to predict customer churn.
• 💬 Restaurant Name Generator: Built an LLM app with LangChain + Streamlit to generate creative, cuisine-specific brand ideas.
• 🚗 Reinforcement Learning for Self-Driving Agents: Designed evolutionary agents via genetic algorithms and neural architecture tuning.

Certifications

• AWS Certified Cloud Practitioner
• Machine Learning
• Deep Learning specializations
• AI Engineer

⭐️ Exploring the boundary between model interpretability, personalization, one custom layer at a time. Hit me up at any platform if you'd like to connect!