Leading Mechanical Supplier (LMS)
Leveraged Technologies
- Python
Platform Integration
- SharePoint
- Mistral LLM (on-prem)
- UKG API
- AWS EC2
*We are using a fictional name and logo forthis case study at the request of the client.
WORK WITH USWe Supercharged Resume Processing For Ai For Mechanical Supplier
Seisan worked directly with key stakeholders to create a functional requirements document that defined scoring logic, job-specific criteria, integration requirements, and performance expectations powered by scalable cloud compute and local AI intelligence.
Objective
Develop a system that can reliably ingest, extract, score, and rank applicants and generate a consistent, detailed report.
Solution
Seisan delivered a customizable automated resume gathering and extraction engine that ingests large volumes of applicant data, analyzes each resume against job- specific criteria, and generates a ranked candidate report.
A leading mechanical supplier specializing in plumbing and HVAC distribution, sought a way to efficiently manage a growing volume of job applicants across multiple locations. Their HR team routinely received hundreds of resumes per day, each requiring consistent evaluation, scoring, and ranking. Seisan developed a fully automated, backend-driven resume screening system capable of processing high-volume applicant data while aligning with strict business rules, on-premises data requirements, and the company’s existing HR ecosystem.
Project Overview
The Company’s HR department was overwhelmed by the sheer number of applicants flowing in daily. Traditional manual review processes created delays, inconsistencies, and bottlenecks—especially when roles required highly specific technical experience.
Key challenges included:
- High-volume applicant intake: Hundreds of resumes daily required rapid overnight processing.
- Inconsistent manual scoring: Subjective review methods created scoring variability across reviewers.
- Dynamic evaluation: Each job posting required its own customized set of screening questions.
- Data privacy requirements: APR required an on-premises LLM, ruling out cloud- based AI solutions.
- Complex integrations: The solution had to interact seamlessly with SharePoint, UKG (APR’s HR platform), and an LLM running locally.
APR needed a system that could reliably ingest, extract, score, and rank applicants—with zero UI requirements—while generating a consistent, detailed report for hiring managers every morning.
During development, Seisan identified performance limitations in the initial LLM provider. The team pivoted to a higher-performing model that ensured overnight processing could be completed reliably without impacting scoring accuracy.
Additionally, inconsistencies between UKG’s public documentation and actual API behavior required iterative testing and custom handling to ensure seamless integration.
Core capabilities included:
- Automated resume collection from SharePoint and other sources
- Dynamic scoring engine driven by screening questions tailored to each job posting
- AI-powered evaluation using a locally hosted Mistral LLM for data-secure content analysis
- Applicant ranking based on job-specific scoring rules Overnight batch processing of all applicants
- Automated report generation summarizing top-ranked candidates
The backend service was deployed in AWS EC2 while maintaining full compliance with APR’s requirement for on-premises AI processing.
Quality Assurance & Readiness
To ensure reliability and accuracy, Seisan executed:
- Unit testing of all processing components
- Regression testing during LLM transition
- Client-driven UAT to validate scoring rules and output consistency
Once validated, the system was deployed and integrated into APR’s nightly hiring workflow.
Results
The Company’s HR team now benefits from a dramatically streamlined applicant screening process:
- Hundreds of resumes processed automatically each night
- Consistent, objective scoring across all roles and applicants
- Faster shortlisting, enabling managers to engage top talent sooner
- Secure, on-premises AI processing aligned with APR’s data policies
- Repeatable, scalable evaluation logic that adapts to new job postings
The system eliminated manual screening bottlenecks and provided APR with a reliable, technology-driven hiring advantage— improving efficiency while preserving full control over applicant scoring criteria.
BACK TO CASE STUDIES