JEANNIE DUARTE

The platform's scoring layer carries two paradigms, both designed and specified here with the company's psychometricians: first, ratings applied at the score level, with the founding psychologist; then the conversion of all scoring and calibration to percentile ratings, with the R&D lead. The Overall and Potential algorithms are of this design, as is the logic for reading percentiles against rating ranges — and an original multi-pass percentile rescoring method in which composite percentiles are re-normed against the benchmark.

Rating adjustment is data-driven and composable: progressive triggers keyed to percentile ranges, with compounding parameters that adjust dependent ratings and emit outputs for recruiting decisions. Configured, deterministic, auditable. The bridge from the psychometric team's statistical analysis platform into production was built here as well — the Excel-structured export format, the import requirements, and the database import routine.

The reporting layer is composable and client-configurable. Elements are mixed and matched by trait type and rendering model, then stacked sequentially: custom trait labels, custom items such as problem-solving scenarios, custom components, and custom reports. Narrative comes from a prioritized-trait comment mechanism, where priority traits are measured against a candidate's scores and database-configured comments are returned in priority order. It has been in production for more than ten years.

Design authority for every user surface — candidate, client, and administrative — across the legacy platform and the redevelopment alike. The fully documented instance is the redevelopment's candidate experience, researched and written here: a designer's primer on selection regulation and applicant-reactions theory, a surface-by-surface design direction across eight surfaces, and working HTML mockups, to WCAG 2.2 AA, with client co-branding carried as a constrained theming layer rather than per-client forks. What makes any of it assessment design rather than web design is that several interaction patterns are locked by measurement validity — the forced-choice direction flip, the timer on a speeded cognitive instrument, the four-point Likert with no midpoint — so the design moves only where moving it cannot change what the instrument measures.

Sole responsibility for the delivery function since 2022. An offshore development team implements to written specification; an infrastructure vendor operates Azure. Everything from the specification through the release sits here.

Scope, specification, and design come first, then handoff with expectations set at handoff. Priority and sequence for the team's work follow, and the project board with them — every JIRA ticket, feature and bug alike, personally written. Stand-ups have run twice weekly since the launch into Azure, often three or four times weekly, and carried the migration from score-based to percentile-based ratings. Follow-up runs through implementation, pre-release QA, and release sign-off; testing of features, fixes, and releases sits here as well.

Vendor staff rotate off the project as they move to other roles, so documentation and ticket discipline are what carry continuity across the handovers. Technical authority over what ships sits here too: the company's owner carries domain authority and is not technical, and the platform's original builders are gone. That concentration of knowledge was named as the first risk in the modernization proposal written for the owner.

Data-preservation and retention analysis through product decommission is led here, and this is the named technical contact for clients' integration shutdown and data-export requests.

Chief API architect for client integrations — project-driven integrations with Workday, iCIMS, and other ATS/HRIS platforms; a bridge integration for a client HRIS that offered no integration surface; and the design and specification of the general API that clients consume.

When the company's I/O staff were gone, client analytics came here directly: score data extracted and shaped for Tableau and AI consumption, analytical workbooks built, and richly formatted data interpretation delivered to clients in person. Data interpretation, not psychometric analysis — the instrument stays with the psychologists.

The platform-modernization decision was designed for a non-technical owner. The vendor's single estimate came bracketed with conservative and aggressive alternatives and a plain-language decision framework; scope was held at the application layer, MongoDB retained. Azure infrastructure was specified as a complete resource bill of materials, and migration sequenced by blast radius — candidates, then clients, then administrators — with staged cutover and rollback. Score-correctness acceptance testing was routed to psychometricians rather than developers, so the people who know whether a score is correct maintain those test cases.

hapbi, the company's internal analytics and operations site, was built solo, human-directed and AI-assisted from concept to implementation. It generates and exports billing, manages pricing with change history, reports on clients and activity, and carries a candidate right-to-be-forgotten module with dry-run preview, audit trail, and response letter. Node/Express on Azure, Entra ID authentication, SQL Server and MongoDB Atlas, CI/CD with staging-slot swaps.

An AI-assisted consultant-summary system was carried to proof of concept in 2023–24, and never formalized or adopted. Prompts assembled the client's job description, the platform report, the client's stated preferences, and the candidate's score data into a draft consultant summary, and every draft was reviewed by the owner before release. It was used on real candidates to cover a consultant absence, and most drafts were released without change. It was later extended to executive assessments with the owner's own rules and writing samples.