Anuj , Data & AI Engineer
I built the credit platform behind Rs 150+ crore of small-ticket lending: medallion pipelines on Databricks, rule-based credit scoring, and a multi-agent AI assistant that lets credit teams query customer data in plain language. Three years in data engineering, now turning raw fintech data into automated lending decisions.
About Me
BIO
It started with a B.Tech in Artificial Intelligence at Vivekananda Global University, where I got my hands dirty with data structures, machine learning, and Hadoop. But the real education began when I stepped into Mandelbulb Technologies as a Data Engineer, building end-to-end pipelines on Azure and Databricks and learning what it actually means to make data reliable in production.
That foundation led me deeper. I started integrating LLMs and open-source AI into production workflows, picked up the DP-600 Microsoft certification, and earned an Agentic AI certificate from DeepLearning.AI. Along the way, I learned that the hardest engineering problems aren't about the tools; they're about understanding how a business actually works and encoding that logic into systems that make lending decisions every day without breaking.
Now at Bharat Fincare, I built CreditDesk: a lending data platform with Databricks medallion pipelines, rule-based credit scoring, an RBI Account Aggregator module for bank statement analysis, and a multi-agent AI assistant for credit teams. The kind of infrastructure that turns a fintech from manual to machine-driven. I'm looking for a team that treats data infrastructure as a competitive advantage, not an afterthought.
DETAILS
Skills & Technologies
Languages
Core LanguagesCloud & Data
Cloud & Data PlatformAI & ML
AI & Machine LearningDevOps & Tools
Infrastructure & ToolingProjects
Client Projects
CreditDesk: Credit Risk & Customer Intelligence Platform
Credit teams were juggling 10 CRM portals, spreadsheets, and manual statement reviews. I built CreditDesk: Databricks medallion pipelines ingesting data from 10 lending products, rule-based behavior scoring, an RBI Account Aggregator module for bank statement analysis, and a multi-agent AI assistant so managers can query customers in plain language. The platform sits behind Rs 150+ crore of disbursed loans.
Fleet Telemetry Real-Time Pipeline
Fleet operations relied on batch processing that delivered insights hours too late. I engineered a streaming pipeline with Azure Event Hubs and Spark transformations orchestrated through ADF into Azure SQL data marts, ingesting 1M+ daily telemetry events and cutting data delay from hours to near-real-time for the dashboards driving fleet cost decisions.
Enterprise RAG Assistant
Teams needed instant access to scattered internal knowledge. I built an AI Assistant using Retrieval-Augmented Generation and Azure OpenAI with Microsoft Fabric OneLake for storage, KQL lookups in ~0.1s, end-to-end LLM responses in 1-2s, and CI/CD on Azure App Service.
Credit Bureau Data Caching Platform
Repetitive CIBIL data fetches via TransUnion were driving high recurring API costs. I built a caching system that extracts key credit data (PAN, name, score) from local sources, stores it in PostgreSQL with AWS S3 backup, significantly cutting third-party API costs and response times.
Credit Analyzer Desktop Application
Credit managers lacked a unified tool to assess risk efficiently. I designed a Credit Analyzer with a PyQt6 desktop client backed by a REST API service, streamlining credit evaluation workflows and improving decision accuracy.
Executive Power BI Dashboard
C-level executives lacked real-time visibility into key business KPIs. I created an interactive Power BI dashboard that identified underperformance areas and provided data-driven recommendations, aligning metrics with organizational goals and decision-making needs.
Personal Projects
Multi-Agent Book Writer (CrewAI)
Writing a structured book requires research, drafting, and editing, typically weeks of sequential human effort. I orchestrated a multi-agent system using CrewAI where specialized AI agents collaborate on each phase, producing cohesive long-form content in under an hour.
Lead Vision Pro: Lead Analytics
Sales teams needed a smarter way to prioritize leads. I built an analytics platform that scores and ranks leads using data-driven signals, helping teams focus on high-conversion prospects and improve pipeline efficiency.
Customer Insights Analytics Platform
Businesses struggle to extract actionable patterns from customer data. I built a Python-based analytics platform that segments customers, identifies behavior trends, and surfaces insights for retention and growth strategies.
Australian Company Data ETL Pipeline
Raw government registration data arrived in inconsistent formats blocking downstream analysis. I built an automated ETL pipeline with data quality validation gates, transforming unreliable source data into clean, analysis-ready datasets.
Bharat Fincare Dashboard
Built a financial reporting dashboard for Bharat Fincare, providing real-time visibility into credit operations, loan processing metrics, and team performance, enabling data-driven management decisions.
Spam Email Classifier (ML)
Built a machine learning classifier to automatically distinguish spam emails from legitimate ones using text features and NLP techniques, demonstrating end-to-end ML pipeline from data preprocessing to model evaluation.
Experience
Data & AI Engineer
Bharat Fincare- Built CreditDesk, the data and AI platform behind Rs 150+ Cr of small-ticket lending across 10 products (60k+ loans, ~30k customers)
- Architected Delta medallion pipelines on Databricks with schema-drift guards and row-level quarantine; rule-based behavior scoring served from PostgreSQL
- Shipped an RBI Account Aggregator module (consent flow, statement analysis) and a multi-agent NL2SQL assistant on Azure AI Foundry for credit teams
Software Developer
Freelancing- Redesigned and enhanced a CSR website using AI-assisted tools (HTML, CSS, JavaScript), improving user experience and performance
- Handled data processing and reporting for CSR campaigns alongside the web work
- Assisted the organization in securing a Google Ad Grant for CSR initiatives, collaborating with Business Development Executives and SEO specialists
Data Engineer
Mandelbulb Technologies- Developed and optimized end-to-end data pipelines using Microsoft Azure, ADF and Databricks
- Improved query performance and data quality, reducing pipeline errors and load times
- Integrated AI solutions with open-source LLMs and OpenAI to enhance data-driven decision-making
Get in Touch
Let's Build Something
I'm looking for Data & AI Engineer roles. If your team needs someone who has already shipped production fintech platforms and real-time pipelines, let's talk.