Soomin Myung

<ul><li>Served as the sole data scientist to architect and deploy the organization’s first unified stock data foundation and integrated operational systems, establishing mission-critical capabilities that did not previously exist. </li><li>Contributed to a £3.9M revenue uplift (£78M turnover company) and saved 5,200+ staff hours annually by institutionalizing governed, end-to-end systems; these replaced fragmented manual processes with accurate, data-driven operations and enabled proactive business planning and response. </li><li>Data Foundation: Achieved 100% automation of stock data recording by engineering a unified platform that eliminated manual processes. Built a PySpark-Parquet ETL engine to ingest 6 years of 41M+ historical records, saving 140+ annual staff hours through a SQL agent scheduler.</li><li>Forecasting: Designed a dual-scenario stock time-series forecasting framework (SARIMA vs. Momentum) to bypass distorted sales signals, enabling reliable item-level out-of-stock risk assessment tailored for 4-month import lead times.</li><li>Unified Operational Layer: Architected a unified operational layer for 30+ cross-functional staff across purchasing, inventory, finance, retail, and logistics using SAP T-SQL and VBA, establishing a single source of truth that eliminated data discrepancies and enabled high-integrity reporting for C-suite decision-making.</li><li>Anomaly Detection: Engineered anomaly-detection frameworks that identified systemic SAP COGS calculation failures and irregular membership-claim patterns (£700K) reviewed by the CEO/CFO.</li><li>AI Automation: Designed an LLM-driven multi-agent purchasing workflow that integrates stock forecasting outputs and supplier & product history to generate structured analysis reports, internal purchasing request documents, and automated supplier-facing drafts, scaling consistent and context-aware decision-making, reducing the team's manual workload by 50% and saving 2,800 staff hours annually.</li></ul>

<p>Python (PySpark), SAP Business One, Parquet, SQL, LLM (RAG/Multi-agent), Tableau, n8n, SARIMA, Excel VBA, Data Quality Management<br>
Served as the sole data scientist to architect and deploy the organization’s first unified stock data foundation and integrated operational systems, establishing mission-critical capabilities that did not previously exist.<br>
Contributed to a £3.9M revenue uplift (£78M turnover company) and saved 5,200+ staff hours annually by institutionalizing governed, end-to-end systems; these replaced fragmented manual processes with accurate, data-driven operations and enabled proactive business planning and response.<br>
Achieved 100% automation of stock data recording by engineering a unified platform that eliminated manual processes.<br>
Built a PySpark-Parquet ETL engine to ingest 6 years of 41M+ historical records, saving 140+ annual staff hours through a SQL agent scheduler.<br>
Designed a dual-scenario stock time-series forecasting framework (SARIMA vs. Momentum) to bypass distorted sales signals, enabling reliable item-level out-of-stock risk assessment tailored for 4-month import lead times.<br>
Architected a unified operational layer for 30+ cross-functional staff across purchasing, inventory, finance, retail, and logistics using SAP T-SQL and VBA, establishing a single source of truth that eliminated data discrepancies and enabled high-integrity reporting for C-suite decision-making.<br>
Engineered anomaly-detection frameworks that identified systemic SAP COGS calculation failures and irregular membership-claim patterns (£700K) reviewed by the CEO/CFO.<br>
Designed an LLM-driven multi-agent purchasing workflow that integrates stock forecasting outputs and supplier & product history to generate structured analysis reports, internal purchasing request documents, and automated supplier-facing drafts, scaling consistent and context-aware decision-making, reducing the team's manual workload by 50% and saving 2,800 staff hours annually.</p>

<ul><li>Quantified the economic impact of a road-to-rail modal shift, estimating £350M–£600M in potential annual public sector savings by modelling reductions in transport externalities such as accidents, noise, and congestion.</li><li>Developed a multi-factor weighting model to adapt European Commission’s transport cost standards to the Middle Eastern market, integrating socio-economic indicators including labour productivity, healthcare development, and population density.</li><li>Synthesised diverse global datasets from the World Bank, Statista, and OECD to build a robust cost-benefit analysis for a large-scale infrastructure feasibility project.</li></ul><p></p>

<p>Statistical Modelling, Data Synthesis, Cost-Benefit Analysis, Impact Quantification<br>
Quantified the economic impact of a road-to-rail modal shift, estimating £350M–£600M in potential annual public sector savings by modelling reductions in transport externalities such as accidents, noise, and congestion.<br>
Developed a multi-factor weighting model to adapt European Commission’s transport cost standards to the Middle Eastern market, integrating socio-economic indicators including labour productivity, healthcare development, and population density.<br>
Synthesised diverse global datasets from the World Bank, Statista, and OECD to build a robust cost-benefit analysis for a large-scale infrastructure feasibility project.</p>

<ul><li>ERP Data Management, Structured Reporting, Stakeholder Management, Data Quality Control</li><li>Managed ERP-based data governance for national energy R&D programmes, ensuring the integrity of project evaluation datasets used for government-level decision-making.</li><li>Standardised complex evaluation reporting processes by synthesising fragmented project data into structured summaries, improving the consistency and transparency of review committee assessments.</li></ul><p></p>

<p>ERP Data Management, Structured Reporting, Stakeholder Management, Data Quality Control<br>
Managed ERP-based data governance for national energy R&D programmes, ensuring the integrity of project evaluation datasets used for government-level decision-making.<br>
Standardised complex evaluation reporting processes by synthesising fragmented project data into structured summaries, improving the consistency and transparency of review committee assessments.</p>

<ul><li>Supported Korea–Uzbekistan e-Government cooperation through statistical reporting and bilateral data exchange. Updated and maintained policy communication pages using HTML and structured content management.</li></ul><p></p>

<ul><li>VBA Automation, Financial Data Analysis, Data Validation, ETL (Excel-based)</li><li>Optimised financial data workflows by developing Excel-VBA tools to automate the extraction and validation of large-scale expenditure data from ERP systems.</li><li>Streamlined financial reporting cycles, reducing manual errors and improving the reliability of budget planning through automated validation scripts.</li></ul><p></p>

<p>VBA Automation, Financial Data Analysis, Data Validation, ETL (Excel-based)<br>
Optimised financial data workflows by developing Excel-VBA tools to automate the extraction and validation of large-scale expenditure data from ERP systems.<br>
Streamlined financial reporting cycles, reducing manual errors and improving the reliability of budget planning through automated validation scripts.</p>