I'm a software engineer and AI systems builder based in Los Angeles. I studied Computer Science at UCLA, where I did machine learning research across two labs: the Julian McAuley Lab at UC San Diego and the Doheny Image Analysis Laboratory, with papers accepted into conferences. You can see them on Google Scholar. I also worked at Google, Amazon, and TikTok, shipping production systems at scale.
I also co-run two businesses: WinTone, a B2B company helping Chinese brands expand into the U.S. retail market, and Keystone Consulting, a college admissions firm with a 100% client success rate. Both gave me a firsthand look at how much time small businesses waste on tasks that AI can handle, so I started building the systems that fix that.
I'm not a consultant who has only read about AI. I've deployed it in businesses I actually run. When I work with you, you get someone who understands both the technology and what it's like to be on the receiving end of it.
I've worked directly with many small businesses and found the same patterns everywhere: time lost to meetings no one recaps, inboxes nobody owns, and reports built manually every week. I approach every engagement from two angles: what the customer experiences and what the internal team actually has to do, then apply what I've seen building systems at Google, Amazon, and TikTok to design something that fits. No off-the-shelf tools. No generic advice. I sit down with you, understand your workflow, and build the thing that actually fixes it.
Summary: Team aligned on pushing the Acme deal to June. Legal needs to review the revised contract by Friday.
Action items:
· Jason — send revised contract to legal by Apr 18
· Sara — follow up with Acme on new timeline
· All — reconvene Apr 22 for status update
● High — Vendor invoice overdue · Reply requested by EOD
● Low — Newsletter from Shopify · No action needed
● High — Client re: contract revision · Approval needed before Friday
Units sold: 1,240 | Revenue: $38,400 | Return rate: 2.1%
Top SKU: Blue Hoodie (312 units)
Report sent to client automatically every Monday at 8 AM.
Step 1: Transcript pulled from Meet, summarized by AI
Step 2: Summary + action items saved to Drive folder
Step 3: Action items emailed to attendees via Gmail
Zero manual steps. Runs automatically every time.
With a distributed team across time zones, emails were getting lost and nothing had a clear owner. I built a centralized system that aggregates all employee inboxes into a single pipeline. Each thread is summarized by AI, tagged with priority, and turned into a shared action list. Bottlenecks that used to surface in weekly calls now surface the same day.
Also automated the weekly client reporting process: performance data pulled, formatted, and emailed to clients without anyone touching a spreadsheet.
Advisors were spending hours a week writing session notes, chasing document submissions, and manually following up with families. I automated the intake and document workflow so application materials move through the pipeline without manual handoffs, and built a meeting summary system so every session ends with a structured recap automatically sent to the client.
The result: advisors spend more time coaching, and no follow-up falls through the cracks.
Tell me what's slowing your team down. I'll be honest about what AI can fix and what it actually takes to build it right.