eReference — Electronic Reference Management Platform
DataScience9 eReference is a cloud-native, multi-role SaaS platform that digitizes and automates the entire reference collection lifecycle — from candidate submission and referee invitation through AI-powered sentiment scoring, fraud detection, and advanced analytics — purpose-built for higher education and enterprise hiring workflows.
End-to-end reference lifecycle
eReference automates the entire reference collection process — from candidate submission and referee invitation through questionnaire completion, sentiment scoring, and final reporting — eliminating manual follow-up.
AI-powered sentiment analysis
Stanford CoreNLP scores every text answer for tone and quality. Lukewarm or negative responses are automatically flagged for human review, surfacing hidden signals that raw text alone would miss.
Multi-layer fraud detection
Device fingerprinting, IP geolocation with VPN/Tor detection, and disposable-email domain validation work in concert to generate a composite fraud risk score and alert HR to suspicious submissions.
Implementation onboarding flow
Discovery
Configure SSO/OIDC provider, import existing candidate and position data from your ATS, set up email and SMS delivery services, and define your initial questionnaire templates.
Integration
Connect ATS APIs for live data sync, configure IP geolocation and VPN detection services, tune email domain allowlists/blocklists, and calibrate Stanford CoreNLP sentiment thresholds.
Data migration
Import historical reference data, migrate existing questionnaire definitions, seed notification templates, and validate fraud detection baselines against historical submissions.
Pilot launch
Run parallel operation alongside existing processes. Conduct HR staff user acceptance testing, complete the WCAG 2.2 AA accessibility audit, and fine-tune fraud detection calibration.
Full go-live
Onboard all departments, decommission legacy processes, activate full monitoring and alerting, and train staff on the analytics dashboard and export tools.
eReference platform scope
Detailed functionality mapped to higher education and enterprise hiring requirements
The core workflow manages reference requests from initiation through completion, with automatic status transitions at each milestone so HR always knows exactly where every request stands.
- HR creates a request by selecting a candidate, open position, questionnaire, required referee count, and deadline. Status auto-sets to pending.
- Status advances automatically: pending → in_progress → completed → cancelled as referees are assigned and submit responses.
- Requests with existing submissions are locked for editing; only the status dropdown remains active to prevent data inconsistency.
- Server-side pagination, keyword search, and status filtering manage large request volumes.
- Full PDF and CSV export respecting all active filters for downstream reporting and compliance.
Built for higher education and enterprise HR
eReference replaces manual email-based reference collection with one auditable, fraud-resistant SaaS platform — WCAG 2.2 AA compliant and ready for ATS integration out of the box.
Talk to our teamPlatform stack
- React 18 + Tailwind CSS — mobile-first, WCAG 2.2 AA compliant
- Spring Boot 4.* / Spring Security / Hibernate JPA
- PostgreSQL 15 (Cloud SQL) — 13-table schema, 5 analytical views
- Stanford CoreNLP — server-side sentiment analysis pipeline
- FingerprintJS + IP Geolocation + email domain validation
- Apache Kafka 3.6 — async notification fan-out
- Google Cloud Platform (GKE Autopilot, Cloud Armor, CDN)
- Apache PDFBox / OpenPDF / Commons CSV — export engine
Key capabilities
- Create reference requests linking candidate, open position, and questionnaire in one workflow.
- Auto-generate time-limited secure access tokens — referees complete questionnaires without creating an account.
- Dispatch invitation and reminder emails/SMS from configurable templates with variable substitution.
- Four answer types: free text, 1–5 star rating, boolean, and multiple choice.
- Asynchronous Stanford CoreNLP sentiment scoring with per-answer requires_review flag.
- Composite fraud risk score aggregating device fingerprint, IP, and email domain signals.
- Real-time dashboard: open requests, pending assignments, overdue deadlines, and sentiment review queue.
- PDF and CSV export for all modules with active filter preservation.
Security and compliance
- AES-256 encryption at rest (Cloud SQL + GCS) and TLS 1.2/1.3 in transit for all connections.
- Endpoint-level RBAC with JWT bearer tokens and TOTP multi-factor authentication.
- SOC 2 Type II, ISO 27001, and PCI DSS compliance inherited from GCP infrastructure.
- FedRAMP Medium/Moderate pathway — all data stored exclusively in GCP us-west2 region.
- WCAG 2.2 AA: screen reader support, keyboard navigation, ARIA labels, and high-contrast mode.
- Referee tokens scoped to a single assignment — cannot access any other record or admin function.
- SSO-ready: OIDC integration with Shibboleth, Okta, and Google Workspace via UWID mapping.