Senior Data Engineer → AI Engineer|Built AI SAAS@RealMadrid Python,LLm,RAG,Agentic AI, Databricks Ex-@Deloitte@Honewell

github

Divyansh Shrivastava  • Peerlist

<p>Built and shipped an end-to-end AI SaaS tool during my Real Madrid internship — integrating social media signals, commercial performance data, and match data into a scenario simulator with LLM-generated insights for senior club decision-makers. -Databricks Gold-layer analytics platform built on unified domain data products — social media (YouTube, Reddit) × historical commercial data × match performance; structured for executive-level decision-making. - LLM-powered scenario simulator built in React — non-technical leadership can model strategic decisions and receive AI-generated narrative analysis without SQL or data access. - Delivered as a standalone SaaS tool to professional football department standards; Real Madrid business team as primary stakeholder.</p>

ClubOS — AI SaaS Platform | Real Madrid Inter

->This project consists of setting up a multinode cluster using Ambari on VMware/Amazon Elastic Compute Cloud2 with two Slave nodes and one Master node as our primary setup. ->We further set up an EDW using the Hadoop ecosystem and ran the full ETL of the data. ->We then analyzed the data through Business Intelligence tools to extract meaningful insights from the data that could be of value to the end-user.

Enterprise Data Warehouse and Data Analysis u

->In this project, we will process the metadata datasets coming with the thousands of records of patients from various hospitals. ->The purpose of publishing this dataset is to figure out the patients depending upon various constraints like the type of disease, depending upon gender, age, seasonal diseases, depending upon location etc.

health care data processing using hadoop eco 

Production-grade football scouting platform covering 15,000+ players across Europe's top 5 leagues — role-based scoring engine, player similarity engine, and an AI Scout Assistant that answers natural language recruitment queries via the Anthropic API. -13 positional profiles scored across 20+ per-90 metrics (xG, xA, progressive carries, PPDA, pressures) — benchmarked by position and league; automated nightly Airflow pipeline with deduplication across 5 data providers - AI Scout Assistant: NL-to-SQL pipeline powered by Claude API — recruitment staff query 15,000+ player database in plain English; validated SELECT-only execution layer prevents destructive queries -Full-stack: FastAPI backend · PostgreSQL + DuckDB · React/Vite frontend · Dockerised deployment · CI pipeline · open-source on GitHub. Tech: Python · FastAPI · PostgreSQL · DuckDB · Anthropic API · RAG · NL-to-SQL · React · Airflow · Docker · dbt

ScoutIQ — Football Scouting Intelligence Plat

Production portfolio project demonstrating end-to-end DE architecture across GCP, AWS, and Databricks — ingesting NYC and London public datasets through a full lakehouse stack with automated data quality and dbt transformation. -5 API integrations (NYC 311, NYPD crime, TfL London transit, AirNow EPA, Open-Meteo); BaseIngester with retry logic, BigQuery streaming loader, S3Writer, 4 Airflow DAGs - dbt transformation layer with automated quality test suite (33+ unit tests), ADR documentation, and lineage tracking -Fully open-source on GitHub — structured as a professional portfolio project with reproducible Docker deployment Tech: GCP (BigQuery, Pub/Sub, Dataflow) · AWS (S3) · Databricks · dbt · Apache Airflow · Python · Docker