My Work
A closer look at products I have built, shown through the screens themselves. Pick a company to browse the work. Short write-ups on each screen are on the way.
The Admin Panel behind dealership onboarding
The internal platform I helped shape for onboarding dealerships, wiring up third-party partner systems, and managing the rules and data that drove the product. A tour through the admin experience.
Admin Panel Homepage
The command center for a dealership account. From a single screen an admin can move a store from onboarding into live operation, and the controls are deliberately simple so nothing high-stakes is triggered by accident. Day Counter is the go-live switch: turning it on clears all test data and starts the dealership on clean, real operational data, so the team never has to separate demo records from real deals. Calculation reruns trade values and prediction scores across the entire inventory on demand, keeping every vehicle current with the latest market data. Deals Sell Status resets the store's deals in one move, giving admins a clean slate without pulling in engineering.
Calculation
This is where pricing logic is tailored to how each store actually sells. A dealer sets distinct rules for Lease, Finance, and Cash buyers and controls exactly how each price is shown to the sales associate on the floor, so the number a rep quotes always matches the store's strategy. Those controls apply separately to Retention and Conquest customers, letting a dealer treat a loyal owner differently from a first-time prospect. Tradebook Region ties valuations to local market data, so trade figures reflect the region the dealership operates in rather than a national average.
Dealer Preferences
A self-serve control panel for shaping the platform around each store. Dealers choose their region and toggle exactly which services are live, from service appointment notifications and CRM sync to OEM inventory and credit score integrations. Because every integration is opt-in, a store only takes on the connections it can support, which keeps onboarding light and the customer-facing experience clean. Dealers also decide which price appears against each vehicle, so inventory always shows the number that fits their pricing strategy.
IP Settings
Larger dealer groups run many locations and departments behind different networks, so access cannot be one-size-fits-all. This screen maps specific IP addresses to specific departments, giving each team the right access without exposing the rest of the platform. It is a quiet but important control that keeps a multi-rooftop operation secure and organized as it grows.
ROI Reports
This is where marketing spend becomes a number a GM can trust. The report shows the ROI of each monthly campaign, comparing performance before and after it went out and attributing closed deals back to the campaign that drove them. Admins can pull any past date range to see which campaigns produced which sales, so the conversation shifts from whether the marketing worked to which campaigns worked, and by how much.
Manage Filters
Clean reporting depends on clean data, and this is where that is enforced. Filters configured during onboarding remove wholesale brokers and other dealerships from closed deals and ROI reporting, so the numbers reflect real retail customers rather than noise. The payoff is that every downstream metric a dealer sees, especially ROI, measures the business they actually care about.
Campaign Filters
Budget control sits directly in the dealer's hands. A store sets how much it wants to spend on campaigns, so marketing scales to the store's appetite rather than a fixed plan. Paired with the ROI reporting, it closes the loop: the dealer sets the spend, then sees exactly what that spend returned.
Mission Control, from zero to one
Mission Cloud's first SaaS product, which I led. It gave AWS enterprise customers real-time visibility into cloud spend, automated checks, and recommendations they could act on.
Mission Control Dashboard
The homepage leads with the customer's Mission Cloud Score, the fastest read on the health of their AWS environment. The score is AI-driven, powered by a proprietary algorithm that adapts to each customer rather than grading everyone against a fixed rubric, so the number reflects that customer's own architecture, usage, and priorities. The model behind it was built, trained, and deployed on AWS SageMaker, and it recalculates dynamically as the environment shifts day to day. Putting it front and center gives customers a reason to check in every morning and turns a sprawling cloud footprint into one figure a business owner can actually track, anchoring every deeper action on the platform to a metric that ties to their bottom line.
Automated Checks
Behind the score sits a library of automated checks mapped to the AWS Well-Architected best-practice pillars. Mission Control continuously evaluates the customer's environment against these standards and takes action on them, so improving security, reliability, and cost is not a manual audit someone has to remember to run. For the customer, it means their infrastructure is held to AWS's own bar around the clock, and their budget is protected by the same engine that moves the score.
Recommendation
This is the customer's action queue. Every recommendation the platform surfaces is listed with when it was last updated and a count of how many actions it involves, so a customer can see at a glance where the biggest opportunities are and choose what to tackle first. Rather than handing customers a static report, it gives them a prioritized, always-current list they can work through, which is what turns insight into realized savings.
Recommendation Drill Down
Selecting a recommendation opens the full picture: exactly what the platform is advising and how many resources across the environment it affects. From this view the customer can accept the recommendation and let Mission Control take action, so the path from understanding a change to executing it is a single step rather than a ticket to another team. Showing the blast radius up front also builds the trust customers need to approve automated change.
Resource Inventory
A complete inventory of every service the customer has deployed, with a toggle that narrows the view to the ones Mission Cloud actively manages. Everything under Managed was monitored by the Mission Cloud Operations team and fed directly into the Mission Cloud Score, so customers can see precisely which parts of their environment the platform stands behind and how that coverage ties back to the number on their dashboard. It makes the scope of the service tangible and connects the score to the real resources driving it.
Clinical Transcript Navigator
A side project I designed and built end to end, outside of any role on my resume, to prove out an AI workflow for healthcare specialists.
After every patient call, a Healthcare Specialist extracts clinical facts from unstructured transcripts, cross-references them against SOPs, and routes the case to the right next step. As volume scales, that manual work creates two compounding problems: specialists lose confidence that no SOP flag was missed, and bottlenecks delay a patient's path to care.
The Clinical Transcript Navigator automates the two highest-friction steps in that workflow, fact extraction and SOP-driven routing, returning a structured summary and prioritized next steps in under 30 seconds so the specialist can focus on judgment, not data entry.
I scoped the problem, designed the workflow, and built the working prototype myself. It is live, so you can run a transcript through it and see the output for yourself.