Deepak Kushwaha
Data Scientist · ML Engineer · Naples, Italy
About Me
Data Scientist with hands-on industry experience building production systems, developing machine learning solutions, and delivering data-driven results across technology and financial sectors. Proven ability to work across the full data lifecycle, from data engineering and model development to deployment and monitoring in cloud environments.
A strong collaborator and team leader, with experience in Agile environments and cross-functional teams. Currently pursuing a Master's in Data Science at the University of Naples Federico II, bringing fresh academic knowledge directly applicable to real-world business challenges.
Recognized for leadership and innovation, having led Team Hindustani to 1st place in the STEM category at Google Challenge Campania 2026, delivering a generative AI solution under competitive conditions.
Projects
Featured / Production Systems
Phishing Detection API
Containerized production API with real-time and batch network security predictions. Full MLOps pipeline with experiment tracking.
View on GitHubVoice2Query
Convert natural speech commands into SQL queries with AI. Features automatic speech recognition, error correction, and real-time visualization.
View on GitHubJob Application Automation
Self-hosted AI pipeline that discovers, scores & tailors job applications automatically. Runs at ~$1.80/month.
View on GitHubDeep Learning & NLP
Churn Modelling (ANN)
Customer churn prediction using Artificial Neural Network with dropout regularization, deployed via Streamlit.
Live DemoImage Classification (CNN)
Multi-class image classifier using CNN with data augmentation and transfer learning techniques.
Architecture Comparison
Image classification comparing CNN (99.08% acc) against baseline Perceptron and ANN on the MNIST dataset.
Next Word Prediction
LSTM-based deep learning model to predict next words in text sequences using NLTK data, deployed via Streamlit.
Word Embedding Pipeline
Foundational NLP pipeline exploring text vectorization and dense representations using Keras Embedding layers.
Machine Learning & Analytics
Movie Recommendation System
Content-based and collaborative filtering movie recommender built with cosine similarity, deployed as an interactive web app.
Live DemoEnd-to-End ML Pipeline
Modular pipeline architecture covering data ingestion, transformation, model training, and prediction with production-grade structure.
View on GitHubBank Marketing Prediction
Classification pipeline predicting subscription outcomes using Random Forest + PCA with hyperparameter tuning.
View on GitHubRegression + Regularization
Ridge, Lasso, Elastic Net with bias-variance trade-off visualization and model comparison.
View on GitHubCustomer Segmentation
Unsupervised clustering with KMeans to identify distinct customer groups for targeted marketing strategies.
View on GitHubSVM Study (Vapnik)
Support Vector Machine theory and implementation exploring kernel methods and maximum-margin classification.
View on GitHubBig Data & Streaming
Spark + Kafka Pipeline
Real-time sentiment and fault classification on high-volume data streams with distributed processing.
View on GitHubExperimentation
Mobile Game A/B Testing
Statistical analysis of A/B test results for Cookie Cats mobile game to evaluate the impact of gate levels on player retention.
View on GitHubTechnical Skills
Machine Learning & Deep Learning
Big Data & Streaming
MLOps & Deployment
Languages & Tools
Professional Journey
MSc Data Science
Data Analyst
Research Intern
Intern
BTech Civil Engineering
Awards
Google Challenge Campania 2026 — 1st Place, STEM
Led Team Hindustani (4 members) at University of Naples Federico II. Built a generative AI solution using Gemini + NotebookLM Pro. Awarded by Google.