Looking for a job

Choi Harim

@choiharim

Message

Production ML Engineer at Hanmac Group who improved PQ bid-award KPI by 35% with a probabilistic prediction system.

South Korea

Message

At Hanmac Group, I built a probabilistic PQ bid prediction system across eight competitive contexts. I improved the group's PQ bid-award KPI by 35% and reached 90.73% chronological coverage.

I also developed procurement NLP models for eligibility and business-category classification. Using weak supervision and teacher–student distillation, I reached 0.9639 eligibility Macro-F1 and deployed CPU inference at about 50 ms per notice.

At Sejong Bid Institute, I built models and Monte Carlo simulations to predict bidder counts and winning prices. I automated training, calibration, batch inference, and AWS deployment, improving the bid-award KPI by 30% over the previous quantile model.

My research and projects include WarpQuant low-bit quantization, weakly supervised segmentation, and retrieval tools. I also developed OCR, AutoCAD add-ins, and ERP integrations at Sanha General Technology, and used OpenCV to automate product-image data processing at Nielsen Korea.

Experience

Work history, roles, and key accomplishments

Researcher

Sejong Bid Institute

Apr 2026 - Aug 2026 (4 months)

Built probabilistic ML and MLOps systems for public procurement, including TabICLv2 and AutoGluon models, Monte Carlo simulations, and automated deployment on AWS with Airflow, S3, and EC2. Improved bid-award KPI by 30% and reduced processing time from 136 to 25 seconds.

Probabilistic Modeling MLOps AutoGluon TabPFN TabM RealMLP Monte Carlo Airflow AWS S3 EC2 Python SQL Docker fastAPI Batch Inference

ML Engineer / Researcher

Hanmac Group

Nov 2024 - Feb 2026 (1 year 3 months)

Developed probabilistic ML, NLP, and RAG systems for public procurement, including PQ bid prediction with R2CCP models, fine-tuned RoBERTa-large for classification, and document retrieval with FAISS and LangGraph. Improved PQ bid-award KPI by 35% and achieved high accuracy in NLP tasks.

Probabilistic Modeling NLP RAG R2CCP RoBERTa LoRAs Weak Supervision Knowledge Distillation Onnx fastAPI Faiss LangGraph Python SQL Airflow AWS

AI / Automation Engineer

Sanha General Technology

Jan 2021 - Nov 2024 (3 years 10 months)

Developed format-specific OCR, AutoCAD add-ins, and ERP integrations to automate document, CAD, and surveying workflows, reducing document work from over four hours to about 30 minutes.

OCR Autocad ERP Automation Python

Data Analyst

Nielsen Korea

Aug 2019 - Oct 2020 (1 year 2 months)

Used OpenCV to extract barcodes and product attributes from images and automated routine processing, reducing the department's product data processing time by about 40%.

OpenCV Image Processing Data Automation Python