I'm a Computer Science student building toward becoming an ML/AI engineer who owns the whole pipeline — not just the modeling notebook. That means preprocessing, feature engineering, training, evaluation, explainability, deployment, and MLOps, treated as one connected system rather than separate exercises.

I care more about systems that could actually ship than notebooks that stop at a metric — which means thinking hard about architecture decisions, failure modes, and how a model behaves once it leaves .ipynb.

• 🧠 CineFusion-X — multimodal deep learning: vision + language + tabular fusion
• 🏗️ MLOps fundamentals — Docker, MLflow, DVC for reproducible, deployable pipelines
• 📚 DSA — strengthening problem-solving & system thinking
• 🔬 Explainable AI — going deeper on interpretability + retrieval systems

Multimodal Movie Intelligence System — reasons about a movie the way a person would, by looking at the poster, reading the plot, and weighing the metadata, instead of relying on a single modality.

The problem — most movie-analytics models lean on tabular metadata alone, throwing away the signal locked inside posters and plot text. CineFusion-X fuses all three into one learned representation.

What makes it interesting — it's not one model, it's a chain of architectural decisions (encoder choice, fusion strategy, latent space design, task balancing) that all have to work together. That systems-level thinking is exactly what I'm optimizing for.

More projects are in progress and will be added as they reach a state worth showing.

I'm always happy to talk about ML systems, architecture decisions, or interesting datasets.

Build real systems. Not just notebooks. Consistency > Motivation.