Curriculum Vitae
I fix the problems that sit between departments.
For fifteen years I've fixed one kind of problem: the work that breaks down between two teams. Each team blames the other, and each is partly right. It happens in operations, marketing, hiring, and AI. This page is organized around that problem, not around job titles.
Outcomes
Every number here comes from the client's own before-and-after reporting, not a formal study. I say that once, here, so I don't repeat it on every card. Open a card to see what changed and why.
30% Operating cost reduction, medical e-commerce Ops
Two failures turned out to be one. A medical supply distributor was treating slow support and frequent stockouts as separate problems and was about to hire for both.
A month of tickets traced to one root cause. Stock state was invisible at the moment a customer asked. So staff guessed and over-ordered.
Response time fell from 24 hours to 2. Stockouts from 30 percent to 5. Operating cost by 30 percent. Neither planned hire happened.
220+ Permanent placements at 90 percent 12-month retention Talent
Measuring recruiters on speed to place quietly selects for the wrong thing. Fast acceptors are not the same as candidates who stay.
Retention became the metric, with automated screening doing the routine filtering underneath. One medical staffing rebuild proved the assumed tradeoff was false: placements ran 50 percent faster and retention still held.
Across these engagements the approach produced 220-plus permanent placements at 90 percent retention at twelve months. The count is cumulative across several years, not a single role.
65% Routine admin calls cut 65 percent, AI assistant for a medical practice AI
The default answer to too many calls is more staff. A clinic drowning in appointment calls and repeat questions was headed exactly there.
Most calls were a few predictable requests: scheduling, reminders, status. An assistant could handle those, but only if it read from current data.
Routine admin calls dropped about 65 percent. No-shows fell close to half. No new headcount. It ran on current data, so it never gave confident answers from old records.
30→5% Stockouts cut with a predictive inventory model AI
An online retailer kept running out of popular items while overstocking slow sellers. Working capital sat in the wrong places.
Reorder calls were made on gut feel and lagging reports. Sales patterns were regular enough to forecast. Nobody had turned that signal into a recommendation.
Stockouts fell from about 30 percent to under 5. A model reading real sales patterns meant reorder amounts were based on real demand, not guesswork.
45% Average ROI across 25+ paid campaigns Growth
No one could say what was working. Paid advertising across client accounts was being bought as a channel rather than run as a system, so spend outran the ability to evaluate it.
Spend was being treated as a cost, not a signal. There was no consistent loop connecting what went out to what came back, so the accounts could not tell what was working until it was too late to act on it.
More than 25 campaigns ran at about 45 percent average return, blended across accounts. Product launches across Amazon, Google, and Walmart were backed by processes that scaled with the traffic.
35% Engagement lift, UNICEF WASH campaigns Nonprofit
Performance gaps between WASH regions kept appearing with no clear cause. The assumption was creative quality. The real reason was structural.
Regional teams were already making good local adaptations. None of it was getting back to central planning. The feedback path did not exist.
Once a return path existed, engagement rose 35 percent and reach 40 percent. The regional differences that were working stayed intact.
15 Countries kept coherent as one coordinated campaign Global
Fifteen countries, each with its own language, channel mix, and local context, coordinated as one campaign.
Coordination failed when it tried to standardize the output. It worked when it standardized the handoffs between regions and the center, and left genuinely local calls local.
Fifteen regional efforts stayed coherent as one campaign. The center never overrode what regions knew about their own audiences.
25% Cost reduction across multi-site clinic operations Healthcare
A pain management practice grew to several locations. Over time each site changed the shared process in its own way. Costs went up and patient experience got uneven.
The divergence lived in the handoffs between scheduling, intake, and billing. Not inside any one team.
Standardizing those handoffs, while leaving local decisions local, cut cost 25 percent with no drop in experience scores.
Capability index
Trajectory
A record of skills acquired, not a list of employers. The company is context. The skill is the part that transfers.
2011 Team leadership and on-site project management AESCO, project manager Southern California +
Ran field projects, crews, and safety programs under regulatory constraint.
Managed geotechnical and environmental field projects end to end. Oversaw crews on site, ran safety meetings, and coordinated across disciplines to keep work compliant and on schedule. The lasting habit came from here: trace the load path before proposing a fix, because the place a thing fails is rarely the place the problem started.
2013 Early-stage startup operations, zero to one FastStart Studio, incubator Irvine, California +
Built and ran operations across several early-stage companies inside a tech incubator.
Worked across several early-stage tech startups inside an Irvine incubator, mostly SaaS and digital-services companies still finding their model. Stood up go-to-market and operational systems from scratch, moved between functions as each company needed, and learned to build processes that could survive the company changing shape underneath them month to month.
2014 Process design and digital operations Sherwood Digital Los Angeles, California +
Made marketing and online-store work run in a clear, repeatable way.
Results depended less on the channel and more on whether the operation behind it could be seen. Digital operations and process work across client accounts, including advertising campaigns and e-commerce launches, made that clear early.
2015 Multi-region program coordination UNICEF WASH, marketing strategy Global program, New York HQ +
Ran one campaign across fifteen countries without erasing what each region knew.
Gaps between regions were usually broken feedback paths, not weak local execution. That was the lesson from advising the WASH communications team on multi-region campaign strategy across fifteen countries. Performance problems looked like creative problems until you traced how regional decisions were or were not getting back to the center.
2016 Process management and technical recruiting CIK Power Distributors Southern California +
Ran industrial process work and built recruiting pipelines at the same time.
Two unrelated settings, one shared problem: work stalling at the boundary between teams who each looked fine on their own numbers. Industrial distribution process and technical staffing pipelines ran at the same time, and the failure mode in each was identical.
2017 Data analytics and machine learning for operational decisions Blackbox AI Remote +
Used ML to filter candidates and find patterns in data. The start of the technical work.
Working from data rather than from assumptions starts here. Applied machine learning to the operational side of recruiting: filtering candidates at scale using pattern-based models, mining data to surface signals that manual review missed. Practitioner ML work on a real problem, not product development.
2019 ML and data science applied across sectors across engagements Remote +
The data and ML work grew past recruiting into operations and forecasting.
Make the data trustworthy first. Check it against reality. Only then make it fast. That order matters. Reverse it and you get fast, confident, wrong answers, which are worse than slow, correct ones. The Blackbox data work grew here into inventory forecasting, process automation, and more. LLMs came later and made the work faster. The foundation was already there.
2020 Multi-site operations and inventory intelligence healthcare and e-commerce clients Remote, US-based clients +
Made clinic operations consistent and gave a live view of inventory.
Rebuilt operations for clinics and e-commerce businesses. Clinic work standardized processes across locations. Inventory work made stock state visible at the moment of the question, rather than adding staff to answer it.
now Independent practice: problems between departments independent practice Los Angeles, California +
Focused on problems that sit between departments, across any sector.
Each team can point to a real failure caused by the other. That exact situation is what this practice focuses on. Independent, based in Los Angeles, working remotely across many fields.
Certifications
Education
Where this helps most
Working across many fields only helps with one kind of problem. Being clear about that is the point. It's what makes a generalist useful instead of stretched too thin.
- A problem sits between two teams, and each one can point to a real failure caused by the other.
- You've already tried a few sensible fixes and none worked. That usually means the problem is being described wrong.
- The same issue keeps showing up in parts of the business that seem unrelated.
- You need someone to find the real problem first, not just another pair of hands to run a plan you already doubt.
- You already know the problem and just need people to do the work.
- The answer depends on specific legal or regulatory rules.
- You need deep expertise in one industry's technical details, the kind that takes ten years to build.