Haoran Jiang

• Developed and configured data pipelines by extracting SaaS data via BICC, staging it in Object Storage, transforming through Data Integration, and loading into ADW to improve data availability and reducing manual data handling. • Supported migration and performance tuning of local warehouse to ADW, optimizing SQL queries and incremental ETL jobs that improved load efficiency and reduced latency for data freshness from D+1 to near-real-time. • Implemented validation scripts and audit checks in OCI Data Integration and Data Catalog to verify data consistency between source files and warehouse tables, greatly improved data accuracy and governance visibility. • Analyzed DNR (Did Not Receive) complaint data and extracted delivery photo evidence to identify submission errors, reduce false positives, improve driver compliance with proof-of-delivery standards, and increase DSP accountability to ensure a 95%+ successful delivery rate. • Conducted root cause analysis of daily delivery failures using SQL and Excel, focusing on key regions, reported findings and action plans during weekly operations meetings, and summarized weekly performance trends in Power BI dashboards to support executive decisions.

• Utilized Python to call APIs for real-time and historical minute-level stock data, conducted trend analysis based on volume indicators and historical data comparisons to identify stocks with clear upward trends. • Developed dynamic visualizations of candlestick charts for potential stocks, displaying real-time stock prices and trading volumes, enhancing the team's trading efficiency and accuracy. • Applied autoregressive (AR) models and linear regression analysis to process the financial data of target stocks, assess operational status, and predict stock price trends. • Calculated key financial ratios (e.g. margin, liquidity, turnover) from multi-period statements to support performance evaluation. • Built dynamic pivot tables to compare financial metrics across years and business units for insight-driven reporting.

• Applied the daily integration and standardization of over 50,000 records across advertising, inventory, and sales datasets, creating a unified data model to support operational and regional reporting. • Built automated KPI pipelines in Python to track advertising ROI, inventory turnover, and fulfillment lead time, enabling near real-time visibility for weekly planning. • Developed a stock-out forecasting model using daily inventory logs, proactively flagging low-stock SKUs to enhance warehouse staging and replenishment across regions. • Conducted pre/post-change performance comparisons via Tablue to evaluate operational impact and summarized insights in weekly reports for stakeholders.

• Spearheaded the ERP-based integration of steel production, inventory, and regional sales data, achieving end-to-end visibility from raw materials to final delivery. • Employed the Kettle tool for ETL operations within the enterprise data warehouse, enabling efficient data transfer from MySQL to Hadoop and setting the stage for advanced development. • Utilized SQL for data development in Hadoop and Spark within the enterprise data warehouse, conducting data manipulation and table creation pipeline with organizational requirements. • Designed Power Bi dashboards to monitor regional sales performance, production throughput, and inventory utilization, identifying supply-demand mismatches and bottlenecks. • Delivered weekly reports with cost and operational insights that informed city-cluster production schedules and transport route adjustments.

• Utilized Python and the PaddleOCR model by Baidu to extract personal information from applicant ID card photos into programmable text message. • Engaged in fine-tuning model parameters with a focus on the recognition of rare characters in ethnic minority languages, as well as complex names and addresses, to increase text extraction accuracy from 77% to 93%. • Implemented data cleaning, validation, and transformation scripts to standardize address, name, and ID formats, improving data quality and downstream query accuracy by 20%.