Curriculum Vitae

• 8+ years of working experience in the field of Data & AI.
• Strong expertise in designing, developing, integrating, and managing large-scale applications across diverse industries, with a strong focus on Generative AI, Cloud, and Big Data solutions.
• Specialized in Generative AI, Large Language Models (LLMs), Data Science, and Blockchain technologies.
• Hands-on experience with state-of-the-art tools and frameworks including OpenAI APIs, Hugging Face Transformers, LangChain, TensorFlow, PyTorch, and Keras.
• Proficient in developing and deploying ML/GenAI workflows using Azure ML, MLflow, AWS SageMaker, and Vertex AI.
• Experienced with modern big data ecosystems such as Databricks, Cloudera, and Delta Lake architecture.
• Professional experience with distributed computing tools: Apache Spark, Hadoop, Hive, HBase, Sqoop, Flume, and NoSQL databases (MongoDB, Cassandra).
• Expert in cloud platforms: Microsoft Azure, Amazon Web Services (AWS), and Google Cloud Platform (GCP).
• Holder of 10+ international certifications and 100+ MOOCs, Nano degrees, and specialized technical trainings.

Cognizant · Kuala Lumpur, Malaysia

Accenture · Malaysia

ASIT Solutions · Malaysia

Logiciel Cube · Malaysia

Malaysia University of Science & Technology (MUST)

University Technology Malaysia (UTM)

Sir Syed University of Science and Technology

Cognizant · Kuala Lumpur, Malaysia

• Designed and implemented a Retrieval-Augmented Generation (RAG) solution using Azure OpenAI and Azure AI Search to power an intelligent Virtual Agent integrated with ServiceNow.
• System securely retrieves knowledge base articles and purchase data (POs/PRs) from ServiceNow via REST APIs, indexes them as vector embeddings in Azure, and answers user queries through a natural language interface.
• Implemented role-based access control (RBAC) to mirror ServiceNow's user permissions.

Cognizant · Kuala Lumpur, Malaysia

• Designed and implemented an agentic AI system that autonomously reads HR policy documents, answers employee queries, and performs actions like applying leave or checking entitlements.
• Multi-agent architecture (Planner, Retriever, Executor) for intelligent task coordination with memory support.
• Deployable as a chatbot (Slack/MS Teams) or web-based assistant via API integration with HRMS.

Cognizant · Kuala Lumpur, Malaysia

• Planned and executed CIMB Data Migration of Teradata from On-prem to AWS including Databases, Hive Tables, Jupyter Notebooks, Dashboards, ML Models, Stored Procedures, Views, and other artifacts.
• Designed the architecture and initialised all services and pipelines for smooth migration.

Cognizant · Kuala Lumpur, Malaysia

• Built and defined Azure Data Factory ETL pipeline extracting data from multiple structured and unstructured sources, transforming and loading it into a centralised data warehouse.
• Transformed data used in Azure ML Studio for training models to improve airline customer experience and satisfaction.

Cognizant · Kuala Lumpur, Malaysia

• Developed multiple Databricks generic pipelines for processing large volumes of encrypted data — decrypting PII data, filtering per policies, performing ETL, and storing in Data Lake for dashboarding and analysis.

ASIT Solutions · Malaysia

• Over 1 million smart energy meters generate data every 5 minutes to Azure IoT Hub; data processed through Azure Data Lake and Databricks, then moved to SQL Data Warehouse and Cosmos DB.
• Configured MLOps to allow ML models to automatically train and publish endpoints via Azure ML.

ASIT Solutions · Malaysia

• Salesperson uses a mobile app to create leads through voice commands. Azure Cognitive Speech-to-Text (STT) used together with LUIS to recognize intents; lead records pushed to CRM.

ASIT Solutions · Malaysia

• Designed and developed BI2020 — migration of a Teradata data warehouse connected to 200+ data sources and over 350 TB of data to Azure SQL Data Warehouse.
• Data transformation processed on Cloudera for big data environment using Spark, HBase, and Hive.

Logiciel Cube · Malaysia

• 50+ TB of emails transferred into Hadoop commodity hardware in real time, converted into sequence files (5× compression), recovering search results within 7–10 seconds.
• ML algorithms applied to generate predictions from the archived data.

Logiciel Cube · Malaysia

• Moved petabytes of call centre voice recordings into Hadoop, processed audio files, converted speech to text, and applied ML algorithms to predict customer engagement with operators.

Python, Power BI, TensorFlow, Tableau, Alteryx, Django, Databricks

SQL Server, IBM DB2, MySQL

Delta Lake Architecture, Big Data, Structured Streaming

AWS Data & AI Workshop · Data Migration Approach Workshop

Terraform, Kafka, Kubernetes, Docker, MLOps