Experience

Summary

Two roles at Microsoft TPD

Cloud Solution Architect — ISV Success

Technical onboarding and architecture support for APAC ISVs listing and selling on the Microsoft Commercial Marketplace

• Ran 10+ Architecture Design and Review Sessions (ADS/ARS) per month; reviewed 60+ ISV Azure architectures in six months
• Advised on offer listing, transactable readiness, and Azure architecture validation
• Advised on packaging strategy by delivery model (SaaS / Managed Application)
• ADS = redesign an on-premises solution into a Marketplace-ready Azure architecture; ARS = validate compliance and deployment readiness

Stack: App Service, Functions, AKS, Container Apps, VMs, Azure SQL, PostgreSQL, Cosmos DB, Blob Storage, Entra ID, RBAC, Key Vault, Front Door, WAF, Managed Applications, ARM templates

Partner Technical Consultant — Azure Data & AI

Sole technical owner of 1:1 presales and deployment consulting for Microsoft AI Cloud Partner Program partners

• Owned the full lifecycle: intake → scoping → research → session delivery → monitoring → closure
• 20+ engagements per month, 120+ in six months, average customer and partner satisfaction 4.75 / 5.0
• Guided 20+ partners from design to production deployment and resolved 60+ technical blockers
• Hardened six partner solutions to production-ready posture across security, cost, and performance

Stack: Microsoft Foundry, Copilot, Copilot Studio, Azure OpenAI, Azure AI Services, Microsoft Fabric, Data Factory, Synapse, Databricks, Cosmos DB, PostgreSQL, ADLS Gen2, Purview, Power Automate, Dataverse, Power BI

Data platform projects

PPAS backup & recovery / data lakehouse

KT

Cut KT's PostgreSQL (PPAS) backup storage from 150TB to 7.53TB (about 95%) with Parquet on HDFS organized as Apache Iceberg tables.

Problem

Full PostgreSQL (PPAS) backups consumed 150TB.

Design

Extracted with Sqoop, stored as Parquet on HDFS, and organized as Apache Iceberg tables. Tiered backups into hourly, daily, weekly, and monthly retention to balance retention and cost.

Rationale

Chose Iceberg over plain Hive tables for snapshot isolation and schema evolution, so history never has to be rebuilt when source schemas change.

Result

Backup storage cut from 150TB to 7.53TB (about 95%). Columnar Parquet improved recovery speed and storage efficiency.

Stack

Apache Iceberg, Parquet, HDFS, Sqoop, Hive, PostgreSQL

Real-time analytics platform enhancement

KT Skylife (7.5M households)

Moved KT Skylife's real-time analytics platform for 7.5M households from CentOS 7 to a Rocky Linux 8 HA architecture with zero downtime and a 20-node scale-out.

Problem

A real-time analytics platform serving 7.5M households ran on end-of-life CentOS 7.

Design

Migrated to a Rocky Linux 8 HA architecture and scaled out 20 distributed servers. Kafka/Flume ingestion → Hadoop/Spark processing → ELK monitoring.

Rationale

RHEL binary compatibility let existing Hadoop/Kafka builds run unchanged; a node-by-node cutover held downtime at zero.

Result

Zero-downtime OS migration with a 20-node scale-out, plus ELK dashboards for real-time data and logs.

Stack

Kafka, Flume, Hadoop, Spark, ELK Stack, Rocky Linux 8

MartDB generative AI data agent (GPT-4o)

KT

A GPT-4o PoC for KT that answers natural-language questions over the PostgreSQL MartDB with RAG and text-to-SQL, cutting query turnaround from hours to seconds.

Problem

Business stakeholders could not query MartDB without SQL, and analysts spent hours on manual queries.

Design

Connected PostgreSQL MartDB to an LLM on Azure AI. RAG plus text-to-SQL answers natural-language questions grounded in MartDB. Streamlit PoC interface.

Rationale

Chose text-to-SQL over pre-aggregation to keep answers current with the operational mart, and grounded generation with RAG to reduce hallucination on schema terms.

Result

Validated LLM querying over the operational mart. Query turnaround from hours to seconds.

Stack

Azure AI Studio, GPT-4o, PostgreSQL, Streamlit, Python

Airflow DAG batch scheduling

KT

Moved KT's Crontab-based Spark and stored-procedure batches to Airflow DAGs on Docker with explicit task dependencies and web-based monitoring.

Problem

Crontab-based batches had no dependency control or visibility.

Design

Analyzed Spark, stored-procedure, and Crontab jobs and moved them to Airflow DAGs with explicit task dependencies, on Docker, with web-based execution monitoring.

Result

Per-task failure tracking, web-based log and schedule inspection, and better maintainability on Docker.

Stack

Apache Airflow, Docker, Spark, PostgreSQL

Insurance system migration from on-premises to AWS

Corea Insurance Solution

Moved an insurer's core services from on-premises Docker hosts to AWS EC2 with security groups, least-privilege IAM, and EBS snapshot backups.

Problem

An insurer's core services ran on on-premises Docker hosts.

Design

Analyzed the full service topology and moved it to AWS EC2/IAM/EBS with security groups, least-privilege IAM, account permission design, and EBS snapshot backups.

Rationale

Kept the existing Docker packaging and moved to EC2 first to shorten the migration, leaving managed-service refactoring for a later phase.

Result

Migration completed with security and permissions in place; EBS snapshots established as the recovery baseline.

Stack

AWS EC2, IAM, EBS, Docker

iOS/Android QA automation

Hana Bank

Automated high-priority cases among Hana Bank's 3,300 manual app test scenarios with Appium and Pytest, running in parallel on three Android and three iOS devices.

Problem

Each banking app release carried 3,300 manual test scenarios.

Design

Automated high-priority cases with Appium and Pytest, running in parallel on three Android and three iOS devices.

Result

Reduced repeated regression effort and verified cross-platform compatibility.

Stack

Appium, Pytest, Python, iOS, Android, Nix-shell

Technical skills