I take over enterprise AI/ML data programs that are behind, over budget, or losing client confidence, and I get them delivering again. I thrive in fast paced environments where requirements change mid-delivery. I work a problem to its actual cause rather than patching the symptom, then build the pipelines, QA systems, and tooling that keep it fixed, so quality becomes a gate before delivery instead of a finding after it. When I see a process that can be improved, I do not wait to be asked. I build the fix and hand it over as something the team can run without me. And when something goes wrong, including my own calls, I say so early, document what happened, and correct it on the record.
The clearest example: two months into my role at Appen, I inherited a natural language to SQL data program for a Fortune 100 enterprise software client that was six months past deadline and $17K past the total contract value, with 92 of roughly 4,000 requested tasks accepted. Our team believed the client kept moving the goalposts. Their team believed we were not fixing anything we were asked to fix. Both sides were arguing in good faith and both were wrong: the pipeline rebuilt its output directory on every run, silently wiping the corrections everyone was making there, so each redelivery shipped the original defect back untouched. The client was ready to terminate. I asked for two weeks and they gave them to me. I patched the pipeline with an edit guard and a delivery preflight, built an automated quality checker that validates and executes every SQL record before delivery, and, with no budget left to buy contributor hours, built an automated query pair generator spanning the required complexity tiers. Twelve days in, two days early, 4,259 QA pairs shipped. No changes were requested and the client accepted the entire batch.
That pattern repeats. On an ASR program for a public safety technology client, I inherited a troubled pilot mid-escalation, rebuilt the correction pipeline with versioned rules and a regression gate, and isolated a systemic timing defect that sat upstream of both the annotators and the pipeline, in the automated segmentation seeds, confirming it acoustically on 27 of 28 files before delivering 63 files and 36,000+ annotated regions with a full change log. On a body camera transcription pilot for a Fortune 100 technology company, I found the inherited rate was roughly a quarter of measured production cost and built the cost models, two pass workflow, and tiered QA taper behind the production proposal.
Before Appen, I owned quality measurement, tracking, and reporting for generative video evaluation programs at Mercor, where I was promoted to Expert Project Manager within 14 days of joining as a domain expert. I designed a consensus scoring method that separated genuine quality failures from tasks where expert disagreement reflected subjectivity rather than error, so ambiguity was no longer scored as a defect. It became the standard across consensus-based projects. I also built a QA dashboard generator that replaced manual weekly metrics reporting and saved each project lead roughly six hours per week; those tools integrated directly with client systems and were used by clients to track quality and performance live.
At Scale AI I earned four promotions in under two years, with scope growing as each process moved from hands-on build to delegated operation: Platinum Team Lead (75 contributors), Senior Coding Quality Manager (30+), Coding Workforce Manager (150+ leads and managers), and Generative AI Project Consultant (1,000+ contributors). As Platinum Team Lead my team held the highest retention, lowest churn, highest activity, and highest quality of any team, with 98%+ contributor engagement and retention. After a company-wide reduction, 96% of the remaining coding team had been hired and trained by me, most advancing into project lead, security engineering, and management roles. I also uncovered $800K in fraudulent payouts from an internal task farm scheme, identified through output volumes no single contributor could produce.
My foundation is 15+ years of management and organizational leadership (2004-2019), including Branch and Regional Loan Office Manager roles and serving as President and Coach of a youth cheerleading league with 125+ participants. After pivoting to technology in 2020, I earned a Summa Cum Laude Bachelor of Science in Computer Science with a concentration in Data Analysis and Project Management in just 27 months while working full-time and raising three daughters with my husband. At SNHU I built 40+ Power BI dashboards across 12 departments, including the sustainability reporting that supported the university's move to Silver status, and discovered a critical vulnerability in a new hybrid cloud data warehouse within 90 days of hire. I am currently pursuing a Master of Science in Engineering in Artificial Intelligence at the University of Pennsylvania.
I work in Python, JavaScript, SQL/T-SQL, and React, across ASR and diarization pipelines, AWS EC2, Firebase, BigQuery, Power BI, and Databricks. What ties it together is that I build for both the technology and the people: quality frameworks for AI training pipelines, tooling that outlives the project it was built for, and contributor systems that reward proven performance. I lead with transparency, and I measure success by the growth of every person on my team.