π§π»βπ»
Alex | Problem Solver & Full-Stack Engineer
10+ Years of Experience β’ HealthcareΒ·PharmaΒ·Insurance Domain Expert Γ AI Agent Lead
"An End-to-End Problem Solver bridging Production Multi-Agent Systems and High-Performance Backends"
π‘ Core Strengths & Value Propositionβ
π§ Agent Orchestration & Tech Management
Extensive experience as a PM breaking down complex projects into micro-tasks and allocating resources effectively. This managing capability seamlessly extends to AI agent development, defining skills/tasks and orchestrating them flawlessly within production workflows.
β‘ High-Performance Architecture
Led the design and auto-scaling of Go (Fiber + Asynq) distributed processing workers, successfully optimizing infrastructure by reducing crawling latency from 7.5 hours to under 5 minutes.
π° Cost & Process Optimization
Focus on maximizing business ROI by eliminating unnecessary cloud costs and optimizing manual processes, reducing maintenance expenses and operational time by over 85%.
𧬠Healthcare & Domain Expertise
Deep domain knowledge in handling large-scale data pipelines and meeting global-level security standards within highly regulated, sensitive industries such as healthcare, pharma, and insurance MyData.
π― Tech Stack (Production Ready)β
AI / ML
LangChainOpenAIGeminiClaudevLLMlocal-llmOCRRAGTorchHuggingFaceWeaviateChromaFull-Stack
TypescriptReactNextjsNuxtjsExpressJSNestJSDartFlutterGoFiberGinAsynqPythonFastAPIDjangoCeleryRedisPostgreSQLMSSQLDevOps / Data
AWSAzureGCPCloudFlareK8sTerraformGitlab CIGithub ActionsArgoCDAirflowDatabricksPowerBISpotfireKeycloakIntroduce.
Starting as a core developer at an early-stage startup, I contributed as a Tech Lead at a leading fintech/insurtech company with over $22M in cumulative funding, and I am currently an IT Architect steering large-scale data pipelines for a global healthcare data platform. Beyond technical completion, I always focus on: "How does this architecture satisfy business targets and optimize team velocity?"
π Hands-on Tech Leadership & E2E Ownership
I am a hands-on leader who directly tackles complex technical bottlenecks, such as system migration strategies and security architecture design. Concurrently, as a Project Manager (PM), I allocate engineering resources efficiently and bridge communication gaps across cross-functional teams to establish an agile development culture that boosts overall productivity.
π€ AI Task Management & Workflow Orchestration
I excel at breaking down vague business requirements into micro-units of skills and tasks. Extending this management philosophy into the AI domain, I isolate the roles of individual agents and orchestrate them into a unified workflow (e.g., LangGraph), building highly reliable and practical Autonomous Data Pipeline environments.
π§ Problem Solving & DevSecOps Automation
I actively eliminate system bottlenecks and human-error-prone points. I reduced operational time by over 85% by establishing automated ETL pipelines for repetitive manual tasks, and further spearheaded the implementation of an enterprise-grade, Zero-Trust-based DevSecOps architecture to drive high-level security-embedded deployment standards.
π Key Business Impactsβ
| Project Domain | Legacy System | Modernized System | Business Impact & Outcome |
|---|---|---|---|
| Healthcare & Patent Data Pipeline | Single-threaded crawler system (Took 17 hours) | Go + Asynq-based distributed scheduling | 99.5% processing time reduction (Within 5 mins) |
| B2B SaaS Platform Transition | Independent on-premise setups (Took 5 months) | Data-Driven Frontend automated SaaS setup | Drastically cut deployment time to under 1 day |
| Insurance Operational Data ETL | Manual work by researchers (2 people, 2+ weeks) | Automated validation and data pipelines | Reduced operational overhead and costs by 85% |
| Insurance AI Recommendation Engine | Legacy combination algorithm with O(N^3) | Conditional filters + ML multi-parameters | Significantly enhanced query response and service quality |
π Highlighted Projectsβ
Enterprise DevSecOps Pipeline & Zero-Trust Cloud Architecture Modernization
π 2026.01 ~ 2026.05
πΌ Cloud Architect & DevSecOps Manager
#GitLab CI/CD #DevSecOps #Azure #Terraform #Policy&Network Configuration #Zero-Trust #Security Audit #Vulnerability #Scan&Pen Test
β νλ‘μ νΈ μμ½
[Problem & Challenge]
To satisfy rigorous global security compliance requirements, I spearheaded vulnerability scanning and penetration testing (Pen-Test) across multiple production environments. The core challenge was remediating infrastructure security risks handling real-time high-volume traffic with zero service downtime.
[Strategy]
Established 'Zero-Trust' principles across all network endpoints and the entire deployment lifecycle. While ensuring high availability during traffic surges, I designed and deployed an enterprise-wide DevSecOps pipeline that seamlessly integrated Terraform-based automated high-speed security scanning (SAST/DAST/IaC).
To satisfy rigorous global security compliance requirements, I spearheaded vulnerability scanning and penetration testing (Pen-Test) across multiple production environments. The core challenge was remediating infrastructure security risks handling real-time high-volume traffic with zero service downtime.
[Strategy]
Established 'Zero-Trust' principles across all network endpoints and the entire deployment lifecycle. While ensuring high availability during traffic surges, I designed and deployed an enterprise-wide DevSecOps pipeline that seamlessly integrated Terraform-based automated high-speed security scanning (SAST/DAST/IaC).
π§ μ£Όμ κΈ°μ¬ & μ±κ³Ό
Zero-Trust Cloud Architecture & High-Traffic Availability
- High-Availability Hub & Spoke Architecture: Adopted a centralized traffic filtering (Azure Firewall) topology, optimizing multi-public IP allocations and auto-scaling policies to completely eliminate SNAT Port Exhaustion risks during heavy traffic spikes.
- Zero-Downtime Zero-Trust Transition: Enforced a 'No Public IP' rule using Azure Application Gateway and Private Endpoints. Implemented phased DNS weighted routing and rolling migrations to completely eliminate impact on live production traffic.
- Tuned Web Application Firewall (WAF) core rule sets to minimize false positives and establish an advanced malicious traffic defense layer.
- Security Isolation & Credential Management: Implemented Azure Key Vault to isolate application secrets from the codebase and constructed a private GitLab Runner ecosystem within a closed network to eliminate internal communication exposure.
- Introduced Terraform (IaC) to prevent human error, ensuring absolute architectural synchronization and versioning across staging and production environments.
DevSecOps Pipeline Optimization & High-Speed Deployment
- Resolving CI/CD Bottlenecks: Enforced strict security scanning (DS, SAST, IaC-SAST, DAST) on every code push, but introduced Incremental Scanning and dependency caching layers to cut pipeline duration, preventing delays in urgent hotfixes.
- Gatekeeper Policy Enforcement: Automated deployment blocking policies based on vulnerability severity to proactively block substandard code from reaching production.
- 3-Tier Deployment Transparency: Streamlined QA continuous deployment (CI) while securing production reliability through a strict manual gatekeepers process requiring Technical Deployment Clearance (TDC/TDA).
- Collaborated with engineering teams to remediate 100% of security vulnerabilities identified via Pen-Testing with zero operational downtime.
Legacy Modernization & High-Performance Distributed Scheduling
- Proactively refactored legacy CoffeeScript (Angular) codebases into high-performance TypeScript React, optimizing frontend rendering performance and maintainability.
- Handled major security version upgrades for NuxtJS and Scala PlayFramework libraries to resolve critical dependency vulnerabilities and maintain backward compatibility.
- Distributed Data Consistency: Migrated legacy PlayFramework AKKA-based scheduling ETLs to a Go Fiber + Asynq distributed architecture. Designed Redis Distributed Locks and an Idempotency guarantee logic to eliminate duplicate processing and data loss risks, achieving 100% data consistency.
π νλ‘μ νΈ ν©νΈ μ±κ³Ό
- Enterprise-Grade Architecture Reliability: Fully satisfied global security compliance audits while maintaining 99.99% infrastructure uptime and availability through optimized load-balancing designs.
- Coexistence of Productivity and Security: Enhanced average development-to-deployment velocity while maintaining strict security controls through caching and incremental pipeline optimizations.
- High-Performance Data Breakthrough: Accelerated ETL data processing throughput and optimized resource utilization by migrating to a Go Fiber distributed scheduling infrastructure with an idempotent architecture.
π‘ νκ³
"This project proved that security and availability are not a trade-off, but foundational components achieved concurrently through advanced architecture." It was a highly meaningful experience of elevating the company's development culture to global standards by leading complex network redesigns under high-volume traffic, zero-downtime migrations, and high-performance distributed systems.
Transitioning Healthcare/Pharmaceutical CRM Platform to SaaS Architecture
π 2025.10 ~ 2026.04
πΌ PM & Full-Stack Tech Lead
#React #NextJS #ExpressJS #PostgreSQL #SaaS #Multi-Tenant #HighAvailability #Redis #RLS
β νλ‘μ νΈ μμ½
[Problem & Challenge]
In the legacy architecture, a dedicated CRM instance had to be built from scratch for each new B2B client, consuming an average setup time of 5 months. Furthermore, consolidating these clients into a single system required a High Availability (HA) architecture to handle large tenant traffic spikes and absolute isolation of sensitive medical data.
[Strategy]
Aimed to eliminate deployment bottlenecks and zero-downtime operational risks through a Backend-Driven configuration approach for common components and high-availability infrastructure scaling. Proposed and led the construction of a Multi-Tenant SaaS platform where secure tenants and custom UI templates can be provisioned in clicks.
In the legacy architecture, a dedicated CRM instance had to be built from scratch for each new B2B client, consuming an average setup time of 5 months. Furthermore, consolidating these clients into a single system required a High Availability (HA) architecture to handle large tenant traffic spikes and absolute isolation of sensitive medical data.
[Strategy]
Aimed to eliminate deployment bottlenecks and zero-downtime operational risks through a Backend-Driven configuration approach for common components and high-availability infrastructure scaling. Proposed and led the construction of a Multi-Tenant SaaS platform where secure tenants and custom UI templates can be provisioned in clicks.
π§ μ£Όμ κΈ°μ¬ & μ±κ³Ό
System Architecture (High-Availability Multi-Tenancy & High-Traffic Design)
- Structured ExpressJS instances to be completely stateless, implementing a horizontal scale-out architecture using load balancers to seamlessly handle user traffic growth.
- Distributed PostgreSQL internal loads by introducing a Primary/Replica replication topology to separate read/write traffic, resolving database bottlenecks via PgBouncer connection pooling.
- Placed a Redis caching layer on top to optimize Backend-Driven configuration lookups, reducing database I/O workloads by over 85%.
- Enforced PostgreSQL RLS (Row-Level Security) at the infrastructure core to protect sensitive healthcare records, fundamentally blocking multi-tenant data leak risks.
Implementation (Traffic Control & Automated Tenant Provisioning)
- Built Express middleware-based tenant-specific Rate Limiting to block sudden traffic surges from individual tenants, protecting shared system resources in the HA environment.
- Implemented an automated provisioning engine within the admin platform that instantly deploys required data schemas, authorization rules, and custom UIs upon organization creation.
- Applied a Zero-Code-based UX so that non-technical personnel in sales and operations can spin up live client demo environments with just one or two clicks.
π νλ‘μ νΈ ν©νΈ μ±κ³Ό
- Drastic CRM Setup Optimization: Reduced the average client provisioning timeline from 5 months to under a single day via end-to-end automation.
- High-Availability Uptime: Ensured robust system availability and prevented outages in massive concurrent request environments through horizontal scaling and distributed database setups.
- Streamlined Infrastructure Maintenance: Consolidated scattered codebases and independent cloud servers into a unified architecture, drastically reducing operational and resource overhead costs.
- Enhanced Operational Agility: Enabled sales and product teams to directly generate and operate live demos without engineering intervention, maximizing cross-functional velocity.
π‘ νκ³
This project went beyond simply reducing engineering hours by migrating fragmented code into an HA SaaS format; it maximized real-world business efficiency by overcoming the inherent architectural limits of multi-tenant environments through database scale-out and data security designs.
Healthcare Data Portal & Integrated Data Engineering Platform Construction
π 2025.10 ~ 2026.05
πΌ Data Engineering & Full-Stack Lead
#Go #Fiber #Asynq #Python #Databricks #Airflow #Celery #ETL Pipeline
β νλ‘μ νΈ μμ½
[Problem & Challenge]
Scattered medical and pharmaceutical public data required manual extraction and processing, resulting in a fragile operational workflow. There were significant risks of queue server degradation and data dropouts during large-scale data ingestion spikes.
[Strategy]
To consolidate separated processes into a unified ecosystem, I built a high-performance distributed crawler using Go and a centralized data pipeline using Airflow, while maximizing pipeline stability through message queue optimizations.
Scattered medical and pharmaceutical public data required manual extraction and processing, resulting in a fragile operational workflow. There were significant risks of queue server degradation and data dropouts during large-scale data ingestion spikes.
[Strategy]
To consolidate separated processes into a unified ecosystem, I built a high-performance distributed crawler using Go and a centralized data pipeline using Airflow, while maximizing pipeline stability through message queue optimizations.
π§ μ£Όμ κΈ°μ¬ & μ±κ³Ό
High-Performance Distributed Parallel Crawler
- Deployed Python (Selenium/Celery) for components requiring dynamic JavaScript parsing, and utilized Go (Fiber) + Asynq workers for core streams where ingestion speed is paramount, maximizing extraction efficiency.
- Zero-Downtime Automated Failover: Overcame Redis container master-slave transition failures by introducing Redis Sentinel and asynq.RedisFailoverClientOption to build an architecture that automatically updates connection pools during topology changes.
- Downtime Minimization via Local Buffering: To handle the 5-10 second timeout while Sentinel promotes a new master, implemented an in-memory ring buffer-based local temporary queue and an exponential backoff retry mechanism inside the Go application, minimizing data loss.
- Preventing Split-Brain and Inconsistency: Configured min-replicas-to-write 1 and min-replicas-max-lag thresholds on the master node to control data loss risks caused by Redis's asynchronous replication, ensuring absolute infrastructure consistency.
- Idempotency Design: Specified a deterministic hash-based unique Task ID derived from the collection URL and target date to guarantee worker-level idempotent processing, preventing duplicate data loading during retries.
- Poison Pill Prevention: Embedded a Panic Recovery middleware and a global Dead Letter Queue (DLQ) limit to prevent infinite retry loops caused by upstream data format mutations.
Airflow Orchestration & Databricks Pipeline Expansion
- Orchestrated the entire lifecycle from ingestion to cleansing, loading, and validation via Airflow DAGs to ensure end-to-end processing transparency.
- Integrated a Databricks-connected data cleansing pipeline within the client infrastructure to satisfy rigorous data quality metrics and metadata policies.
Operational Data Management Portal
- Built an API-driven data management web application allowing internal and external stakeholders to directly search, filter, and extract cleansed data.
π νλ‘μ νΈ ν©νΈ μ±κ³Ό
- Completely automated legacy manual data collection and processing workflows into a hands-off scheduled pipeline.
- Increased pipeline uptime and minimized data drops to near zero through distributed message queue and high-availability Redis optimizations.
- Unified siloed data spaces scattered across multiple internal departments into a single data lake catalog solution.
π‘ νκ³
I implemented a massive data lifecycle end-to-end, spanning from the backend collection layer to big data ETLs and the user-facing web portal. In particular, it was a profound experience to fundamentally resolve failover gaps and data loss vulnerabilities inherent in distributed queue architectures (Asynq/Redis) under high-volume traffic using infrastructure tuning, Go internal buffering, and idempotent software designs.
π Professional Work Historyβ
IQVIA is the world's leading global provider of advanced AI-driven healthcare data, technology solutions, and clinical research services.
Within the technology division, I manage Full-Stack platform development, project management, and AI/Data Engineering infrastructure.
Principal Architect & Tech Lead / Architecture, IT Design & Development
Global IT Audit & DevOps Security Policy Standardization
Cloud Infra & Security Ops
2026.01 ~ 2026.05
- Automated policy compliance by aligning Azure-based internal infrastructure and network CI/CD (GitLab) with strict global security standards.
- Established a serverless cloud cost defense line and dynamic concurrency limits to prevent infrastructure cost spikes during sudden traffic surges.
- Implemented a Zero-Trust Cloud Architecture and secured global infrastructure governance through IaC.
- Built DevSecOps processes, significantly reducing global security audit approval lead times.
- Improved system vulnerabilities and maintained compliance standards through strategic Legacy Code Modernization.
[Lead] Global Healthcare/Pharma CRM Platform SaaS Transition
PM & Full-Stack
2025.10 ~ 2026.04
- Led the migration of isolated client-specific instances into a unified Multi-Tenant SaaS system to deliver a configuration-based independent platform.
- Introduced tenant-specific Rate Limiting and Redis distributed caching to prevent noisy neighbor resource monopolization issues.
- Spearheaded dynamic rendering optimization and reduced CRM provisioning time by applying a Zero-Code admin UI approach.
- Designed and implemented a multi-tenancy architecture to ensure absolute data isolation.
[Lead] Public Data Workflow Integration Platform & API Portal Development
Data Engineering & Full-Stack Lead
2025.10 ~ 2026.05
- Designed and optimized large-scale data pipeline collection and ingestion using Go and Python for automated harvesting.
- Applied backpressure control mechanisms to prevent OOM issues and cope with target site blocking (IP Ban) or data spikes.
- Led the construction of a real-time API portal based on distributed locks and integrated an Airflow & Databricks-driven end-to-end quality check.
- Enhanced data collection throughput and secured portal stability by building a high-performance distributed parallel crawler.
Data Catalog RAG Solution Launch & Operations
AI Engineering & Data Ops
2025.01 ~ 2026.04
- Introduced a PgVector and OpenAI/LangChain-based RAG infrastructure for utilizing internal data/service catalogs.
- Built a semantic caching framework and a FastAPI inference server to overcome LLM API Rate Limit latencies during user spikes.
- Core PgVector scaling: Optimized database indexing and connection pooling to maintain high-performance queries across large-scale vector spaces.
- Connected and managed Spotfire/PowerBI cloud dashboards, linking them to real-time monitoring pipelines.
Distributed Scheduler Backend System Replacement & Optimization
Full-Stack Developer
2024.07 ~ 2025.08
- Drove performance improvements by replacing the core pipeline execution layer with a Go (Fiber + Asynq)-based asynchronous queue environment.
- Secured worker pool tuning and a Dead Letter Queue (DLQ)-based crash recovery mechanism to control queue bottlenecks during task surges.
- Managed in-house infrastructure across local, development, and production servers using Nginx, Jenkins, and KT Cloud, establishing a zero-downtime deployment pipeline.
A premier Korean insurtech B2C/B2B platform enterprise maintaining a 99% customer retention rate and over $220M in cumulative insurance brokerage value, leading to a successful Series C investment round.
Tech Lead / Data Platform Cell
B2B Financial Sector (Securities) Custom Insurance Analysis/Recommendation Engine v4 Development
Backend Engineer
2022.07 ~ 2022.09
- Established an independent deployment architecture by packaging the ExpressJS backend into a single executable binary using PKG to satisfy securities firm security regulations and on-premise/closed-network deployment constraints.
- Designed a schema normalization pipeline using Pandas to handle heterogeneous financial data formats from the standard government MyData system and internal metadata, ensuring data integrity.
- Introduced a Keras-based multi-class deep learning model and processed domain research data with Pandas to highly segment the diagnosis scoring algorithm for securities clients, maximizing recommendation reliability.
- Built a flexible metadata management system using ExpressJS to adapt dynamically to data structure changes, reducing lead times for reflecting custom requirements per financial institution.
Coverage-Level Precision Analysis & API SaaS-Based Insurance Engine v3 Development
PM & Full-Stack Engineer
2022.04 ~ 2022.08
- Transitioned the computation layer to a Golang-based parallel processing structure to secure infrastructure scalability, moving from product-centric analysis to massive granularity coverage-level dense analysis.
- Rebuilt the ExpressJS-based data pipeline and back-office to efficiently serve large metadata and computation results from the Python (Keras, Pandas)-based deep learning engine.
- Newly constructed an integrated admin platform using NestJS and NextJS infrastructure sets to support the transition to a B2B API Service (SaaS) business model, enabling multi-tenancy management.
B2B Simplified Insurance Claim Automation Solution Launch
Backend Engineer
2022.02 ~ 2022.05
- Designed and developed an in-house Go-based PNG coordinate mapping PDF document engine with the goal of minimizing external API dependencies.
- Introduced streaming processing and worker pool-based concurrency control to prevent memory OOM risks during large-scale document conversions.
- Independentized the claim network pipeline from fax/email systems, utilizing a message queue-based asynchronous processing layer with guaranteed idempotency to cut maintenance costs and enhance traffic flexibility.
Bodoc Engine Core Module - Insurance Diagnosis Combination & Product Recommendation (v2)
PM & Full-Stack Engineer
2020.11 ~ 2021.01
- Completely redesigned a high-risk technical debt legacy solution (which failed migration by other teams due to proprietary library dependencies and tight coupling) into a modular, independent engine architecture capable of standalone operation and custom delivery.
- Resolved the ultra-massive combinations problem (2^216 combination explosion) and server crashes during multi-product optimal price combination calculations through an essential coverage pre-filtering (pruning) strategy built in collaboration with domain researchers (compressing the valid candidate pool to 42,738 to secure real-time operational stability).
- Built a data parsing pipeline between ExpressJS and Python backends, directly building a Sklearn-based 3-class multinomial logistic regression model in production and optimizing calculation parameter serving.
- Achieved a 40x improvement in the core engine's average calculation speed by minimizing Core DB communication and separating transaction/log data ingestion layers.
- Newly built an advanced admin platform based on React and ExpressJS dedicated to data verification/processing, automating the manual collection processes of domain researchers and cutting operational time by 85%.
Hybrid App Customer Monitoring / CRM Project
Full-Stack Engineer
2019.11 ~ 2020.03
- Developed a deep link routing architecture and flexible business code capable of handling sudden large-scale traffic explosions (thundering herd) caused by massive targeted marketing push notifications.
- Led React architecture refactoring and asset optimization (code splitting) to prevent memory leaks in the hybrid app webview and optimize user experience.
- Minimized app crash rates under abnormal entry paths and exceptional scenarios by building a client-side Error Boundary.
Executed data infrastructure establishment and core product modernizations at early-stage startups focused on healthcare services and AI messaging systems.
Backend / Data Engineer
Medical Device & Health Examination Result Data Verification and Operational Pipeline Construction
PM & Full-Stack Engineer
2019.08 ~ 2019.10
- Developed a custom rule-based algorithm and integrated management platform to cleanse incomplete, unstructured data parsed via OCR from physical medical devices and audit data reliability.
- Led the end-to-end data architecture design and launch spanning 'Hospital Management System β‘οΈ In-house ETL Pipeline β‘οΈ Customer-facing Mobile Web App' to achieve rapid market validation (MVP) for the early-stage healthcare service.
In-House Core Service Structural Analysis & Legacy Migration Stability Design
Backend Developer
2018.12 ~ 2019.07
- Established a gradual architecture transition strategy centering on the introduction of ReactJS and ExpressJS to minimize crash risks in high-complexity legacy stacks, while analyzing infrastructure bottlenecks.
- Designed and stably released small-scale core prototype services, such as a mass messaging integration app, within a short timeframe to discover new business models.
π Freelancer & Side Projects
8am (B2C & B2B Match Platform)
Lead Freelancer
Customer-Worker Matching Platform Modernization & High-Availability Deployment Pipeline
PM & Full-Stack Architect
2025.08 ~ 2025.10
- Designed a next-generation platform backbone incorporating a NextJS and NestJS-based asynchronous event handling structure to prevent bottlenecks during real-time quote requests and matching traffic surges.
- Introduced Docker container orchestration onto the AWS infrastructure environment to secure sustainable production deployment stability, automating the deployment process to eliminate human error and enhance infrastructure isolation.
Mobile Application Lead
Senior Health Monitoring App Core Logic & Sensor Data Processing Engine Optimization
App Developer
2025.07 ~ 2025.09
- Established a local-remote availability synchronization mobile application architecture that prevents healthcare data loss for senior users via Flutter background thread control.
- Implemented a high-performance computation algorithm pipeline that pre-processes raw acceleration sensor data streams and quantifies them into activity scores to optimize device battery consumption and prevent main UI thread bottlenecks.
Grablo (No-code IoT Controller Builder)
Lead Architect Freelancer
Global Target No-code IoT Platform Cloud Infrastructure Zero-Downtime Stabilization Advising
Lead Architect
2023.12 ~ 2026.01
- Formulated and technically supported a high-risk zero-downtime migration strategy to transition systems into a GKE (Google Kubernetes Engine) environment within Google Cloud Platform (GCP) while maintaining global live service availability.
- Performed advising on backend API layer latency improvements, load balancing optimization, and horizontal pod autoscaling (HPA) architecture to sustain massive IoT device connections and global user requests.
Full-Stack Developer
B2C Gift/Return-Gift Release Service Prototyping & Infrastructure Construction
PM & Full-Stack Engineer
2022.05 ~ 2023.05
- Oversat the entire product lifecycle design from the ideation phase to the cross-platform (Flutter/SvelteJS) mobile implementation release.
- Dedicatedly managed Oracle database schema design, backend integration, and cloud server deployment considering transactional safety to support successful market entry.
Full-Stack Engineer
Legacy Backend API Architecture Optimization & Internal Operations System Integration
PM & Full-Stack Engineer
2022.08 ~ 2022.12
- Refactored exception handling layers of unstructured old legacy source code, elevating NestJS API architecture stability to reduce runtime crashes and establish a Swagger-based automated API specification framework.
- Integrated an in-house data operations dashboard based on low-code tools (Retool) to minimize internal operational overhead and maintain data consistency, improving back-office workflow efficiency.
Frontend Engineer
Global Game Supply & Publisher Operations Admin Web Platform Development Support
Frontend Engineer
2021.08 ~ 2021.12
- Developed and implemented virtualization techniques and high-performance React admin component architectures that prevent browser rendering degradation even when loading massive row data, ensuring visibility for large game publishers.
- Designed a reusable core component library to ensure UI consistency and enhance internal frontend productivity.
π οΈ Skills & Tools (Production Proven)
Listed stacks and architectural practices reflect components directly designed, deployed, and scaled in enterprise production environments.
AI / LLM Engineering
LangGraph β’ LangChain β’ OpenAI/Gemini/Claude β’ Ollama β’ LlamaIndex β’ Torch β’ HuggingFace β’ Transformers β’ QLoRA β’ vLLM
Backend & Distributed Systems
Go (Fiber, Asynq) β’ NodeJS (NestJS, Express, BullMQ) β’ Python (FastAPI, Django, Celery)
Frontend & App
TypeScript (NextJS, Nuxtjs, React, TailwindCSS, Shadcn/UI, React Native) β’ Dart (Flutter)
Data & MLOps
Airflow β’ Databricks β’ Spark β’ Pandas β’ Spotfire β’ PowerBI β’ PostgreSQL β’ Weaviate β’ Chroma β’ Redis β’ Elasticsearch
Cloud & DevOps
GCP β’ AWS β’ Azure β’ Kubernetes β’ Terraform
Summary & Next Stepβ
"An expert senior full-stack architect who designs systems so that architectural optimization efficiency and infrastructure reliability translate into exponential business growth value."
I have consistently redesigned high-risk technical debts into standalone engines capable of B2B delivery and SaaS conversion. I broke through uncomputable domains caused by ultra-massive combinations (2^216 combination explosion) using heuristic pruning strategies to lead algorithm optimization. Even under strict regulations of financial (insurtech/MyData) and healthcare domains alongside on-premise closed-network deployment constraints, I have quantitatively proven that architectural design directly drives business outcomes, achieving a 40x improvement in engine computation speeds and an 85% reduction in internal operational overhead through automation.
I deeply respect single responsibility principle (SRP)-based modularization, data pipeline integrity, and robust container orchestration cultures that maintain system availability without collapsing during massive traffic surges.
Now, I look forward to tackling the new architectural challenge of the LLM era: Agentic Workflows. As a technical leader who quietly solves problems and elevates team capabilities, I am eager to drive tangible business growth together.