Data Engineering · Enterprise Analytics · Applied AI

Turning complex data into intelligent systems.

I'm Anuj Shukla, a data and analytics engineer with 9+ years of experience spanning enterprise BI, cloud data architecture, real-time integration, predictive machine learning, and agentic AI. I build the systems that connect data, decisions, and operations.

Explore my workGet in touchView resume
Explore the tabs above for professional experience, an interactive Power BI report, and guided ADP / TTM demonstrations.
01 / About

Engineering depth. Business clarity.

My work connects enterprise data platforms with the people and processes that rely on them. From modernizing Power BI and Microsoft Fabric ecosystems to architecting AWS streaming pipelines and predictive analytics applications, I focus on dependable data foundations, useful intelligence, and thoughtful delivery.

01 / CLOUD & DATA

Real-time to historical

Streaming, CDC, micro-batch ingestion, governed warehousing, and data reconciliation across enterprise systems.

02 / INTELLIGENCE

Analytics & machine learning

Semantic models, Power BI, operational KPIs, predictive regression, and decision-support services.

03 / DELIVERY

Architecture & leadership

BI modernization, technical planning, cross-functional collaboration, Scrum delivery, and quality-focused engineering.

02 / Work & Experience

Professional experience & projects.

Enterprise business intelligence, cloud data engineering, and predictive analytics — with interactive demonstrations available in the Simulators tab.

ENTERPRISE EXPERIENCE · JUL 2019 – PRESENT

Miracle Software Systems | Client: Americold Logistics

Senior BI Developer · Data Engineer · Scrum Master

Building enterprise BI, custom Power BI experiences, modern cloud analytics, and financial reporting solutions for third-party logistics operations.

Enterprise Power BI · i-3PL Analytics

Designed and delivered operational dashboards, interactive KPI reports, and reusable semantic models for logistics stakeholders. Developed advanced DAX measures, Power Query transformations, dynamic slicers, bookmarks, parameters, buttons, and drill-through experiences; optimized SQL queries, views, functions, data models, and report performance across enterprise data sources.

Custom Default View & Secure Reporting

Conceptualized and developed a reusable Default View feature that personalizes a report's initial KPIs and visuals through username-aware DAX measures. Unlike Row-Level Security (RLS), Default View does not restrict the underlying dataset: users initially see their assigned facility but can select other facilities from the slicer to analyze shared business data. Also implemented RLS where access restrictions were required, and built cross-report navigation that preserves filter context across reports.

BI Modernization · SSRS to Power BI & Microsoft Fabric

Led migration strategy and source mapping for 100+ SSRS reports and the transition from on-premises Power BI Report Server to Power BI Service. Designed Microsoft Fabric solutions with OneLake, Lakehouse, Warehouse, Data Factory pipelines, Bronze/Silver/Gold layers, and star/snowflake semantic models; supported reconciliation, incremental refresh, gateways, workspace access, deployment, and performance tuning.

AR / DSO Analytics & Delivery Leadership

Supported receivables and days-sales-outstanding analytics by designing automated SharePoint/Excel → Power Automate → AWS SFTP → Amazon Redshift data flows. Coordinated requirements, sprint planning, Jira delivery, stakeholder demonstrations, and technical guidance across BI and engineering teams.

Adaptive Data Platform (ADP) · Real-Time Data Engineering & Agentic AI

Architected a warehouse data integration platform connecting WMS and WES operational systems through real-time streaming, micro-batch ingestion, and change data capture (CDC). Built AWS Kinesis, Lambda, and DynamoDB pipelines, combining low-latency operational intelligence with Redshift Serverless historical analytics, reconciliation, and event traceability.

Designed an AI-assisted data-quality orchestration framework with specialist agents for logical duplicates, datetime consistency, reference integrity, workflow sequencing, missing fields, and quantity anomalies. Implemented evidence-based decisions, FIFO-aware processing, durable tracking, controlled replay, and human review for ambiguous exceptions.

Technologies: AWS Kinesis · Lambda · DynamoDB · Redshift Serverless · CDC · Python · Agentic AI

Turn Time Management (TTM) · Predictive Analytics & Machine Learning

Developed a predictive analytics solution to estimate warehouse and trailer turnaround times and support dock planning, resource allocation, and operational SLA decisions. Engineered Python feature pipelines and Gradient Boosting Regression models using operational factors such as queue depth, pallet volumes, dock availability, and processing delays.

Integrated predictions with FastAPI services and PostgreSQL operational records, providing estimated turn times and actionable recommendations. Containerized application and database components with Docker and validated end-to-end inference and operational workflows.

Technologies: Python · scikit-learn · Gradient Boosting Regression · FastAPI · PostgreSQL · Docker

FEATURED PROJECTS

Explore the interactive simulators

Choose a demonstration to see the features and workflows in action.

Power BI · Americold

Interactive BI & Personalized Reporting

Experience custom Default View versus RLS, filter-persistent navigation, bookmarks, parameters, drill-through, and KPIs.

Open Power BI simulator →
Adaptive Data Platform

Real-Time Data Engineering & Agentic AI

Follow warehouse events through streaming ingestion, AI quality checks, and trusted analytics.

Open ADP simulator →
Turn Time Management

Predictive Analytics & Operations

Explore predicted trailer turn times, dock-capacity factors, and operational recommendations.

Open TTM simulator →
DATABASE DEVELOPMENT · MAR 2015 – MAY 2017

Anant Softtech Private Limited

SQL Developer

Developed and maintained SQL queries and database routines supporting point-of-sale applications on SQL Server and Oracle. Worked on data retrieval, reporting queries, data validation, and troubleshooting to support dependable transactional reporting.

03 / SIMULATORS

Explore interactive demonstrations.

Select a project below, or jump here directly from Work & Projects.

Power BIADPTTM
03 / Interactive BI

Power BI feature showcase

Try Default View versus RLS, filter-persistent navigation, bookmarks, drill-through, parameters, and KPIs.

INTERACTIVE POWER BI EXPERIENCE · SYNTHETIC DATA

Warehouse Operations Intelligence

Explore the custom Default View approach, true row-level security, and connected report experiences. Switch users, select facilities, and compare the results.

Sample report
INTERACTIVE REPORT EXPERIENCE

Choose how the report starts

Both modes begin with the signed-in user's assigned facility. The difference is what the user can access afterward.

Sample identity: Alex · Atlanta. You can change the user inside the report.

04 / Data Engineering

Adaptive Data Platform

Explore real-time integration and evidence-based AI quality decisions.

Cloud data engineering / Agentic AI

Adaptive Data Platform (ADP)

An enterprise data integration and intelligence platform combining event-driven pipelines, historical analytics, quality controls, and specialized AI-assisted issue resolution.

Architecture & delivery

  • Designed real-time streaming, scheduled micro-batch, and change data capture ingestion patterns across operational source systems.
  • Used AWS Kinesis, Lambda, DynamoDB, and Redshift Serverless to support low-latency processing and historical analytics.
  • Built operational monitoring, data reconciliation, and event traceability into the data flow.

Reliability & intelligence

  • Developed specialized data-quality agents for logical duplicates, temporal inconsistencies, reference integrity, workflow sequencing, missing fields, and quantity anomalies.
  • Designed evidence-based review and approval paths for ambiguous data issues.
  • Implemented FIFO-aware orchestration, durable decision tracking, and controlled replay and recovery patterns.

Technology stack

AWS KinesisAWS LambdaDynamoDBRedshift ServerlessAmazon S3CDCSQLPythonOllama / local LLMsDocker

Architecture overview · Adaptive Data Platform

Enterprise SourcesWMS · WES · CDC · Batch
→
Streaming & ProcessingKinesis · Lambda
→
Operational StateDynamoDB · Quality Gates
→
AnalyticsRedshift · Governed Metrics

Simplified conceptual architecture; specialized paths support historical and near-real-time processing.

INTERACTIVE SYSTEM WALKTHROUGH · ADP

Follow an event from warehouse source to trusted analytics

Explore how a multi-source data platform processes operational events, checks quality, and handles exceptions before they reach analytics. All events are synthetic.

1 Choose an event

Ready to trace a warehouse event.

Demo of architecture and decision logic; no connection to live AWS services.

2 Trace the data journey
01WMS / WESOperational event
→
02Kinesis / LambdaIngest and normalize
→
03AI quality gateValidate and route
→
04RedshiftTrusted analytics
3 Understand the decision
READY

Choose an event and select Process event to see what happens.

Assigned specialist—
Analytics destination—

Why it matters: the platform protects downstream reporting from incorrect or unverified events.

Explore engineering decisions and technical depth
  • Separated near-real-time operational processing from historical analytical serving to support distinct latency and retention requirements.
  • Designed scoped quality gates for logical duplicates, timestamp consistency, reference integrity, event sequencing, missing fields, and quantity validation.
  • Implemented evidence-based exception handling with durable review and replay paths; ambiguous cases require human approval.
  • Used bounded asynchronous orchestration and FIFO-aware continuation to preserve processing order and improve recovery behavior.
  • Built reusable AI-agent contracts with constrained execution and local LLM integration.
05 / Predictive Analytics

Turn Time Management

Explore predictive warehouse operations and turn-time recommendations.

Predictive analytics / Machine learning

Turn Time Management (TTM)

A predictive operations platform for estimating warehouse and trailer turnaround times, combining machine learning with operational data and application services.

Explore technical architecture & implementation

Predictive analytics

  • Engineered Python-based preparation and feature workflows using structured operational records.
  • Developed Gradient Boosting Regression models to estimate turn times for planning and operational decisions.
  • Analyzed delay patterns, turnaround indicators, and process bottlenecks to support operational insights.

Application engineering

  • Integrated model inference with FastAPI endpoints and PostgreSQL-backed workflows.
  • Connected predictions with operational records to support application-level consumption.
  • Containerized services with Docker and validated end-to-end application integration.

Technology stack

Pythonscikit-learnGradient BoostingFastAPIPostgreSQLDockerREST APIsSQL

Architecture overview · Turn Time Management

Operational DataWarehouse & Trailer Events
→
Feature EngineeringPython · Time & Process Signals
→
PredictionGradient Boosting Regression
→
Application ServicesFastAPI · PostgreSQL

Conceptual flow showing analytics and application integration.

INTERACTIVE APPOINTMENT & PREDICTION DEMO · TTM

Schedule a trailer. Predict its turn time. Improve its SLA outcome.

Create an appointment, review the predicted turn time, then apply operational recommendations to see the potential SLA improvement.

1 · Create appointment2 · Predict performance3 · Apply recommendations4 · Compare SLA
1 Create a warehouse appointment
Adjust operational conditions · Queue, docks, delays & SLA

Step 1: Create an appointment to start the guided simulation.

Synthetic demonstration with a transparent illustrative estimate. Not connected to live operations or the trained Gradient Boosting model.

2 Prediction and SLA — before action
Complete the appointment form and select Create appointment & predict.
Behind the engineering

In the actual TTM project, historical warehouse operations data is prepared for a Gradient Boosting Regression model. Predictions are served through Python-based application services with PostgreSQL-backed operational workflows. This portfolio demonstration illustrates the decision-making experience with synthetic data and a transparent calculation rather than presenting fabricated model outputs.

Explore modeling and operational analytics
  • Analyzed turn-time variability, delay drivers, and warehouse execution patterns to identify operational bottlenecks.
  • Prepared predictive features from process milestones and time-related operational signals.
  • Built regression-based estimates to support forward-looking operational planning.
  • Connected analytics workflows to backend services and persistent relational storage for repeatable use.
  • Structured containerized application components for consistent execution and maintainability.
04 / Expertise

Tools, platforms & practices.

Business intelligence

Power BI, DAX, Power Query, Bookmarks, Parameters, drill-through, Incremental Refresh, RLS, SSRS, semantic models, dimensional modeling.

Cloud & data engineering

AWS Lambda, Kinesis, DynamoDB, Redshift, S3, Glue, Microsoft Fabric, OneLake, Lakehouse, Data Warehouse, Azure Data Factory, ETL/ELT, CDC, streaming.

AI & application engineering

Python, scikit-learn, Gradient Boosting Regression, agentic AI workflows, Ollama, REST APIs, FastAPI, Docker, Git.

Databases & delivery

SQL, PostgreSQL, Oracle, SQL Server, Amazon Redshift, query optimization, data validation, technical architecture, Agile and Scrum.

05 / Education

Academic foundation.

M.S. in Computer Science

New Jersey Institute of Technology · Newark, New Jersey · 2019

B.S. in Information Technology

Gujarat Technological University · Ahmedabad, India · 2017

06 / Contact

Let's build something useful.

Interested in data platforms, enterprise analytics, AI/ML, or technical collaboration? I'd be glad to connect.

Email

anuj.b.shukla@hotmail.com

Phone

+1 (678) 979-0668