Praneeth Varma Danthuluri — Data Analyst & Analytics Engineer

About

From raw data to decisions that ship

I'm a Data Analyst and Analytics Engineer with an MSc in Artificial Intelligence and 4+ years of professional experience — including 2+ years building automated reporting workflows, data models, and dashboards. Power BI, SQL, Python, and Microsoft Fabric are my day-to-day tools, with hands-on workflow experience across dbt, Airflow, and Snowflake.

I've automated manual reporting workflows, kept scheduled daily data refreshes running reliably, and delivered analytics-ready datasets covering tens of thousands of retail transactions — the kind senior stakeholders actually use to make decisions.

What sets my work apart is the full-stack view: I don't just build dashboards — I own the pipeline behind them, from ingestion and modelling through data quality tests, governance, and the final semantic layer stakeholders actually use.

Modern data stack, end to enddbt · Airflow · Fabric · Snowflake · Power BI

Reporting on autopilotDaily refreshes with validation checks — zero manual steps

Governance-minded deliveryData quality tests, integrity controls, audit-ready docs

Available immediatelyIreland · Stamp 1G work authorisation

Skills

The toolkit

Strongest day to day: SQL, Power BI & DAX, Python, and Microsoft Fabric. Hands-on workflow experience with dbt, Airflow, and Snowflake. Highlighted pills are the tools I'd be happy to whiteboard in an interview.

Analytics & Languages

SQLPython (Pandas, NumPy)EDAStatistical AnalysisData Cleaning & ValidationFeature Engineering

BI & Visualisation

Power BIDAXPower Query (M)TableauLooker StudioStreamlitMatplotlib

Data Engineering

Microsoft Fabric (Lakehouse, Data Factory, Semantic Models)ETL Workflow AutomationdbtApache AirflowDimensional ModellingData Warehousing

Databases & Cloud

MySQLSQL ServerSnowflakePySparkAWSAzureDockerGitCI/CD

ML & AI

Machine LearningScikit-learnTensorFlowCNNLSTMTransfer LearningPredictive Modelling

Governance & Delivery

Data GovernanceIntegrity ControlsAudit ReadinessAgile/Scrum

Experience

Where I've delivered

Four years across analytics engineering, reporting, and software delivery — most recently owning the full data stack for a multi-site retail operation.

Data and Operations Analyst

May 2023 – Mar 2026

JKSR Trading LimitedDublin, Ireland

Retail Analytics & Reporting: Supported analytics and operational reporting across three retail locations, working with sales, stock, reconciliation, and performance data
Automated Reporting: Replaced manual weekly reporting with an automated Power BI and Microsoft Fabric solution — cutting reporting effort by ~40% and enabling daily KPI visibility
Analytics Workflows: Built production-style analytics workflows using SQL, Python, dbt-style transformations, scheduled data refreshes, and validation checks for retail sales and inventory reporting

Scheduled Refreshes: Set up daily data refreshes with Apache Airflow — task dependencies, retries, and failure alerts — so dashboards updated without manual intervention
Central Data Store: Consolidated cleaned data into Snowflake, writing SQL queries and views used as the single reference point for downstream Power BI reporting
Forecasting & Trend Prediction: Built forecasting and trend prediction scripts in Python to support demand planning and inventory decisions
Root Cause & Governance: Investigated data issues to identify root causes, maintained data integrity standards and audit-ready documentation
Stakeholder Collaboration: Partnered with operations and business teams to translate requirements into analytical tasks

Project Service Coordinator

Jan 2022 – Dec 2022

Mokshar Creative StudiosHyderabad, India

Used SQL and Excel to translate business requirements into structured reporting frameworks
Maintained project data tracking systems to support traceability and informed decision-making
Prepared reports and visualisations to communicate trends to senior stakeholders

Assistant Software Engineer

Mar 2021 – Dec 2021

Tavant TechnologiesBengaluru, India

Designed RESTful APIs enabling efficient data access and integration across platforms
Implemented CI/CD pipelines and automated testing using Git
Contributed to data-driven product improvements in Agile/Scrum environments

Professional Work

Case study: retail reporting, automated

One production project, end to end — the problem, the decisions behind it, and what changed.

JKSR Trading · 3 retail locations · May 2023 – Mar 2026

Retail Operations Analytics Platform

The problem

Weekly performance reporting for three retail locations was fully manual — exporting sales and stock data, cleaning it in spreadsheets, and rebuilding the same reports every week. Slow, error-prone, and always a week behind the business.

The approach

Consolidated sales, inventory, and reconciliation data into a central store, built layered SQL transformations (staging to reporting) with validation checks for nulls, duplicates, and reconciliation mismatches, scheduled daily refreshes with retries and failure alerts, and modelled the result as a Power BI semantic model with DAX measures in Microsoft Fabric.

A decision worth defending

Chose scheduled daily batch refreshes over near-real-time. Buying, staffing, and stock decisions were made at day level, so daily was the right cost-and-complexity tradeoff — reliability over novelty. Validation checks ran before every refresh so a bad load never silently reached a dashboard.

The outcome

~40% less time spent on reporting, dashboards covering tens of thousands of transaction records refreshed daily with zero manual steps, and one agreed source of truth across all three locations instead of three competing spreadsheets.

SQLPythonPower BIDAXMicrosoft FabricdbtAirflowSnowflake

Internal production system, so the data stays private — happy to walk through the data model, measures, and dashboards in an interview.

Self-Directed Build

Production work does not always expose you to the failure modes interviewers actually probe. I built a second pipeline, outside of work, specifically to hit those cases on purpose — and documented what broke.

Self-Directed · Postgres → Snowflake → dbt → Airflow · Aug 2026

Retail Medallion Pipeline

The problem

JKSR gave me production experience with dbt, Airflow, and Snowflake — but a working pipeline does not always force you through the failure modes that get probed in interviews: late-arriving records, silent schema drift, a task retrying mid-run. I built a pipeline specifically to hit those on purpose.

The approach

Postgres OLTP source seeded with ~9.3M rows of realistic retail sales and inventory data, extracted via a watermark-based Python job into Snowflake, transformed through dbt Core staging → intermediate → marts layers into a star schema with an SCD Type 2 customer dimension and 47 automated tests, orchestrated on Airflow with retries and failure alerting.

A decision worth defending

The incremental model filters on pipeline load time, not the source update timestamp — so a record corrected two years after the fact still merges in cleanly instead of being silently dropped. I proved it by editing an 18-month-old order line and watching it land as one updated row, not a duplicate.

The outcome

9.3M rows loaded in 3m22s; the incremental re-run dropped from 4.5M rows to 56. Renamed a source column on purpose and found the pipeline stayed green while writing nulls — accepted_values ignores nulls without a paired not_null test. Query tuning cut scanned micro-partitions from 8/8 to 1/8.

PostgreSQLSnowflakedbt CoreAirflowPythonDocker

Public repo — every deliberate breakage, the diagnosis, and the fix are documented alongside the code.

View on GitHub

MSc & Research Projects

Machine learning and deep learning work from my MSc in Artificial Intelligence — a separate track from my professional analytics work, kept here for range.

MSc Thesis · NCI Dublin · 2025

Network Intrusion Detection System

Hybrid CNN-LSTM model detecting DDoS, Botnet, and Brute Force attacks in real time — 98.2% F1-score on the CIC-IDS2017 dataset (2.8M+ records). End-to-end Python pipeline with SMOTE for class imbalance; real-time Streamlit monitoring dashboard containerised with Docker.

PythonTensorFlowScikit-learnDockerStreamlit

View on GitHub

Computer Vision · Transfer Learning

Car Damage Classification System

Benchmarked VGG16, ResNet50, and EfficientNet transfer-learning architectures for automated vehicle damage detection — a rigorous, comparative deep-learning evaluation for a real-world insurance-adjacent use case.

PythonTensorFlowCNNsTransfer LearningGit

View on GitHub

ML · Imbalanced Classification

Credit Card Fraud Detection System

Benchmarked Logistic Regression, Random Forest, and Decision Tree classifiers; applied SMOTE and Grid Search with Cross-Validation for robust tuning. Deployed a Flask web interface for real-time fraud prediction.

PythonScikit-learnFlaskPandasJupyter

View on GitHub

Education

Academic foundation

MSc Artificial Intelligence

National College of Ireland, Dublin

2023 – 2025

BTech Computer Science Engineering

Vellore Institute of Technology, India

2017 – 2021

Available Immediately · Stamp 1G · Ireland

Let's build something data-driven

Open to Data Analyst, BI Analyst, Analytics Engineer, and Junior Data Engineer roles in Ireland. Stamp 1G — full-time work authorisation until June 2027, no sponsorship needed to start.

praneethvarmaraju3@gmail.com +353 83 417 7274 LinkedIn GitHub