TSIA
Jun. 2021 – PresentAI Automation Engineer
Feb. 2026 – Present- Started TSIA's internal AI automation initiative across research and content operations, self-initiated and expanded to every process that fit, each built with the validation harnesses, reusable skills, and documentation the team needed to run it.
- Designed and shipped the LLM judging system behind an industry awards program, scoring 131 submissions in the 2026 live cycle. Reason-then-rate architecture across three data-derived personas on hybrid text and rendered page images, with independent models from two other providers as referees.
- Validated it against the human expert panel before it was trusted with decisions: matched the human winner in 7 of 7 categories across 29 held-out applications, with Impact reaching rho 0.71 against a human-to-human ceiling of 0.53. Established that ceiling first, diagnosing an apparent ICC of 0.14 to 0.24 as range restriction.
- Cut judging effort from 330 human grading hours to 10 hours of human-in-the-loop review, a 97% reduction.
- Replaced a 40-step manual poll checklist spanning five systems with a four-phase tool gated by four approvals, 1,246 lines across 16 Python modules. A first-time operator drove the second full cycle end to end with zero manual system edits, returning at least 32 hours a year per analyst.
- Deployed under real governance constraints. The AI pair-programmer never saw member data, working from structure and aggregate diagnostics only, verified against synthetic fixtures under read-only credentials.
- Won TSIA's company-wide AI Innovation and Enablement Award.
Principal Research Analyst
Feb. 2025 – Feb. 2026- Senior individual contributor for the research organization, owning solution design across the benchmark, survey, and conference programs.
- Replaced roughly 8 near-duplicate validation scripts with a single config-driven engine covering all 8 research practices, where adding a practice became a configuration entry instead of new code.
- Built the conference-to-CMS integration: 72 sessions filtered from 110 raw records, with author mapping validated against archived sessions at 58 of 61 exact, every mismatch explained, before any production write.
Data Engineer
Mar. 2022 – Feb. 2025- Led the migration of the firm's financial indices off manual vendor-workbook entry onto an API-driven Alteryx pipeline into BigQuery, with Tableau outlier detection. Owned the quarterly refresh from trainee to pipeline owner, 2022 through 2026.
Data Analyst
Jun. 2021 – Feb. 2022- Gathered requirements and mapped processes with stakeholders across teams. Delivered ad hoc statistical analysis, including a logistic regression on member renewal rates reported as non-significant (p = .128) rather than dressed up as a finding.