Dolly Sahu

Dolly Sahu

AI Engineer & Data Engineer

Building intelligent solutions that solve real-world problems related to Data and AI | Master of AI from Monash | 3+ years industry experience with Accenture, TCS and A8 Consultancy

View My Work ↓ Get in Touch

Trusted Experience With

A8 Consulting
Accenture
Tata Consultancy Services
GradGirls Tech

Areas of Expertise

💾

Data Engineering (BFSI)

  • SAP ECC to S/4HANA migration at TCS
  • Legacy COBOL system to S/4HANA migration at Accenture
  • 14M+ financial records processing with zero data loss
  • 155+ ETL batches monitored & optimized daily
  • 70+ production-grade ETL batch jobs developed with SAP BODS
  • 100+ HANA calculation views maintained & optimized
  • 40+ annual production deployments with full documentation
  • Enterprise data governance for sales, finance & insurance domains
🤖

AI/ML (Insurance | Health | CS)

  • Machine learning eligibility engine for world's largest insurer
  • RAG systems & LLM pipelines for customer support automation
  • Eliminated 30+ hours of manual document lookup weekly
  • Predictive modeling for arthritis risk assessment (76% accuracy)
  • Computer vision attendance system enabling instant class snapshots
  • Advanced statistical modeling for healthcare applications
⚙️

Automation (Insurance & Industrial)

  • Workflow automation for monthly report generation (10+ hours saved)
  • Intelligent document parsing from 35+ employee shared drives
  • Automated invoice processing (handwritten & digital) to PostgreSQL for Transport management system
  • 200+ batch job orchestration using JavaScript & Control-M
  • CI/CD pipeline automation for production deployments
  • Operational excellence through intelligent automation

Featured Projects

A showcase of AI/ML and automation projects demonstrating practical applications that deliver measurable business value

🤖 ApplianceSense RAG
AI Engineer @ A8 Consulting
Production-grade RAG pipeline automating document lookup across 1,500+ technical manuals for warranty support agents.
✨ 80% reduction in manual document lookup time | 75% accuracy achieved
LangChain ChromaDB Azure OpenAI OCR Python
The Problem
Warranty agents manually searching 1,500+ PDFs, spending 30+ hours weekly on document lookup
What I built and how
OCR pipeline + ChromaDB semantic search + Azure OpenAI GPT-4 for natural language responses
Key impact
80% lookup time reduction | 75% accuracy | 30+ hours weekly saved
💊 ArthritEase Platform
AI Engineer Lead @ Monash
Personalized health platform for arthritis patients with conversational AI assistant and weather-based pain management.
✨ 76% accuracy predictive modeling | Real-time weather integration | Serverless architecture
Generative AI AWS Lambda Weather APIs ML Algorithms
The Problem
Existing Arthritis apps give generic information without considering real-time situation, pain level, demography, or lifestyle
What I built and how
Conversational AI assistant + weather API integration + predictive ML model on AWS Lambda
Key impact
76% prediction accuracy | Real-time weather insights | Serverless architecture
📊 SAP S/4HANA Migration
Data Engineer @ Accenture
Large-scale digital transformation migrating legacy COBOL systems to SAP S/4HANA for global insurance leader.
✨ 40% data retrieval improvement | 35% decision accuracy increase | 40% manual review reduction
SAP BODS XGBoost HANA Python
The Problem
Legacy COBOL systems slow; 14M+ financial records need migration to S/4HANA
What I built and how
SAP BODS ETL pipelines + XGBoost ML models + 155 batch job orchestration
Key impact
40% data retrieval improvement | 35% accuracy increase | Zero data loss
🚗 Autonomous Racing Car
Autonomous Software Engineer @ Monash
Computer vision and ML replacement for LIDAR systems on driverless racing car M23, using ROS tools.
✨ Vision-based perception system | Real-time processing | Motion control integration
ROS OpenCV Python Computer Vision
The Problem
LIDAR systems expensive, inaccurate during rain, need CV-based alternative for real-time perception
What I built and how
Computer vision pipeline with OpenCV + ML models + ROS integration for motion control
Key impact
Real-time perception | Low-latency processing | Seamless ROS integration
🏠 SmartAttendance
ML Engineer @ College
Automated attendance system using face recognition to eliminate manual roll calls in educational settings.
✨ Real-time detection | Instant attendance logging | Zero manual intervention
Python OpenCV ML Face Detection
The Problem
Manual attendance rollcalls time-consuming and prone to proxy in classroom
What I built and how
Face detection using OpenCV + Python ML pipeline + instant database logging
Key impact
Real-time detection | Instant logging | Zero manual intervention
📄 Invoice Automation
Current Project 2026
End-to-end automation converting handwritten invoices to database entries using OCR and image processing.
✨ Image to Text conversion | Automated DB insertion | Zero manual data entry
Python GCP Node.js PostgreSQL
The Problem
Manual invoice data entry from handwritten/scanned docs; high error rates and slow processing
What I built and how
OCR + image processing + Python automation + PostgreSQL for intelligent data extraction
Key impact
Image to text conversion | Automated DB insertion | Zero manual data entry

Skills I Gained by Projects

A comprehensive toolkit spanning AI/ML frameworks, data engineering, cloud platforms, and modern development practices

🤖 AI/ML & LLMs

RAG LangChain PyTorch TensorFlow Scikit-learn XGBoost

💾 Data Engineering

SAP HANA SAP BODS ETL/ELT SQL Python Pandas

☁️ Cloud & DevOps

AWS Lambda GCP Azure Jenkins CI/CD Docker

🛠️ Development Tools

Git JavaScript Node.js Control-M MLflow

🎯 Computer Vision & NLP

OpenCV OCR YOLO Mask R-CNN NLP

📊 Methodologies

Agile/Scrum Data Governance Technical Docs Domain-Driven Design

Experience & Achievements

A track record of driving measurable impact across AI, data engineering, and automation

AI Engineer
A8 Consulting Pty Ltd, Melbourne
October 2025 – Present
Building production-grade RAG pipelines and AI solutions for enterprise clients.
  • End-to-End RAG Pipeline: Architected production-grade retrieval system transforming 1,500+ unstructured PDFs into searchable vector database using Tesseract OCR, ChromaDB, and LangChain.
  • LLM Integration: Integrated Azure OpenAI GPT-4 with ChromaDB semantic search for context-aware manual lookup and error-code interpretation across appliance documentation.
  • Data Quality: Optimized retrieval accuracy to 75% through OCR refinement and feature engineering workflows, ensuring high-fidelity LLM inputs.
  • Business Impact: Reduced manual lookup time by 80%, delivering 30+ hours of weekly operational savings by automating support queries.
  • Production Deployment: Transitioned prototype to production; currently under evaluation for accuracy, usability, and workflow integration.
OTD Volunteer
GradGirls - Women 4 STEM, Melbourne
January 2026 – Present
Supporting women-in-tech initiatives partnered with 25+ leading Australian companies including Liberty Financial.
  • Provide on-the-day support to moderate live chat, track attendance, escalate participant questions to speakers or MCs, and handle basic troubleshooting.
  • Supporting event coordination and community building initiatives for a women-in-tech program partnered with 25+ leading Australian companies.
Data Engineer
Accenture, Remote (US)
June 2022 – July 2023
Led digital transformation for global insurance leader, migrating legacy systems to SAP S/4HANA.
  • Engineered complex ETL orchestration mapping COBOL structures to SAP schemas, reducing delays by 40%
  • Built XGBoost Eligibility Engine improving decision accuracy by 35% and reducing reviews by 40%
  • Developed high-concurrency reporting pipeline replacing batch processes with live dashboards
  • Implemented automated PII/SPI masking ensuring strict data governance compliance
Assistant Data Engineer
Tata Consultancy Services, Remote (India)
January 2021 – June 2022
Managed technical migration of TATA Group's sales infrastructure from SAP ECC to S/4HANA.
  • Cut data processing time from 2.5 hours to 40 minutes by optimizing 150+ HANA Calculation Views
  • Orchestrated 30+ error-free production deployments using Jenkins CI/CD pipelines
  • Proactively migrated Talend ETL jobs to BODS, saving 3 months of future workload
  • Trained 4 team members to manage streamlined workflows built in production
Master of Artificial Intelligence
Monash University, Melbourne
July 2023 – August 2025
Advanced studies in machine learning, computer vision, reinforcement learning, and NLP with real-world projects.
  • Led ArthritEase team building health platform with generative AI and 76% accurate prediction models
  • Developed autonomous racing car perception system for driverless vehicle competition
  • Conducted research on AI applications solving real-world problems while adhering to MLOps principles

Get in Touch

I'm always excited to connect with fellow AI enthusiasts, potential collaborators, and employers. Whether it's a project idea, technical discussion, or opportunity, I'd love to hear from you!

Email
[email protected]
Phone
+61 466 133 802
LinkedIn
linkedin.com/in/dollys

I typically respond within 24 hours. For urgent matters, reach out via email or LinkedIn.

Start a Conversation