Anant Singh

what's in the toolbox

Skills & Tools

The stack I keep reaching for, from training a model to shipping a Spring Boot API.

AI/ML

Python (Pandas, NumPy, Scikit-learn) for classification, regression, and recommender systems, with NLP preprocessing — tokenization, stemming, feature extraction. Building toward AI engineering: Generative AI, Retrieval-Augmented Generation (RAG) pipelines, vector databases (Pinecone, FAISS, Chroma) for semantic search, embeddings, and LLM-powered application development.

PythonScikit-learnNLPGenerative AIRAGVector DatabasesLLMsLangChain

Java & Backend Development

Object-oriented programming in Java, building RESTful APIs and microservices with Spring Boot, ORM with Hibernate/JPA, dependency management via Maven, and unit testing with JUnit — the backbone for scalable, production-ready backend systems.

JavaSpring BootREST APIsMicroservicesHibernate/JPAMavenJUnitOOP & Design Patterns

Data & BI

SQL (joins, subqueries, window functions), Power BI (DAX, Power Query, data modeling), and Excel for the quick-and-dirty cleanups.

SQLPower BIDAXExcel

Cloud & GenAI

AWS AI services and Amazon SageMaker, generative AI & prompt engineering, with an eye on responsible, evaluated AI models.

AWSSageMakerGenAIPrompt Eng.

the sketchbook

Academic Projects

Six end-to-end builds — from raw data to a dashboard someone could actually use.

Real-Time Stock Analysis GitHub ↗

A real-time market intelligence system combining live stock data (AAPL, TSLA, AMZN & more) with news APIs, processing 1,000+ data points a day.

AI sentiment pipeline hit ~85–90% accuracy on financial news, lifting signal reliability by 30%.
Automated Python + SQL + LangChain pipeline cut manual analysis time by 60%, under 5-sec latency.
Power BI dashboards for trends, sentiment & correlations sped up decisions by 40%.

Customer Shopping Behaviour Dashboard GitHub ↗

Mined 50,000+ transaction records to uncover purchasing trends and segmentation opportunities.

Improved marketing targeting accuracy by 32% through segmentation analysis.
Built a Power BI dashboard with 15+ DAX measures, cutting manual reporting by 60%.
Surfaced high-value segments behind a 25% lift in campaign conversions.

Instagram User Analytics Dashboard GitHub ↗

End-to-end analysis of 120,000+ engagement records using SQL-driven data modeling.

Optimized PostgreSQL schema with joins & aggregations for growth and engagement trends.
20+ KPI visualizations in Power BI, cutting manual reporting time by 55%.
Found posting patterns behind a simulated 28% engagement lift.

Uber Rides Analysis Dashboard GitHub ↗

Analyzed 100,000+ ride records to uncover demand patterns and peak booking hours.

Identified opportunities for 22% better operational planning.
Cleaned & preprocessed data with Python/SQL, improving reliability by 95%.
Engineered time-based features to optimize ride allocation at peak hours.

Movie Recommender GitHub ↗

Content-based recommender using cosine similarity across 10,000+ movie records.

Combined genres, cast, keywords & overviews into vectorized features, lifting precision by 28%.
Simulated user relevance scores improved by 34%.
Pulled posters & ratings from external movie APIs for a richer UI.

Spam SMS Classifier GitHub ↗

ML-based spam detector tuned on 5,000+ labeled messages, reaching 96% accuracy.

TF-IDF + Naive Bayes / Logistic Regression cut false positives by 18%.
Full NLP preprocessing pipeline: tokenization, stopword removal, stemming.
Shipped with a Tkinter GUI for real-time, click-and-check classification.

the paper trail

Education & Certifications

The credentials backing up the projects above.

🎓 Education

B.Tech, Information Technology

KIET Deemed to be University, Delhi-NCR

September 2024 – June 2028

Building intelligent
systems!
— one model, one API at a time.