A decade turning operational bottlenecks into technology. At every company - real estate, health-tech, programmatic advertising - I find what's manual, build the tools to automate it, and scale the team around it.
Most recently: CTO and Acting COO at JointCommerce, where I hired and led a 15-person team while owning the $300K-$500K/mo P&L. Managed 200+ active accounts every month at 85% gross margins - nearly double the industry average of 45%.
This is the pattern at every company. Build → Automate → Margin → Growth → Repeat.
A cross-section of accounts I've managed, activated, or built measurement for - across regulated industries, Fortune 500 brands, national franchises, and political campaigns.
Two-sided AdTech platform - consumer ordering + advertiser activation, built end-to-end. 15-person startup, $3.6M-6M ARR. I hired every team member, built the technology, ran the P&L, and grew the client base from zero to 200+ simultaneous accounts. 85% client retention with zero long-term contracts.
Built and scaled the consumer ordering platform from 2,000 to 200,000 monthly users in 11 months. 100x growth, end-to-end.
Consumer app deep dive →CDP behind 200+ advertiser accounts: 21M+ unique consumer profiles, 1B+ behavioral data points, 6 DSP integrations - including activation by a major DSP partner for pharmaceutical advertiser campaigns.
CDP deep dive →
Live Interactive Demo
I built this to solve the "Creative Wall" in retail advertising. Instead of a 3-day manual design cycle, this engine generates interactive ad units in under 2 hours directly from client product feeds.
Built the entire technology platform: 235 models, custom pixel tracking, 6 DSP integrations, ad network, and e-commerce. That platform held 85% gross margins and powered 200+ simultaneous accounts.
View Architecture Deep Dive →
Ad format I built: 9 industry presets, 7 templates, A/B testing. Generates interactive shoppable units from product feeds in under 2 hours. 3x CTR vs standard display.
View Full Story →Programmatic advertising platform (~125 employees). #1 performer out of 22. Managed $900K+/month in ad spend across 10+ DSPs, peaking at $1.2M/month during high-volume periods. Built tools that elevated the entire organization - not just my own performance. Contributed to a multimillion-dollar exclusivity deal.
Direct programmatic work with major advertisers across CPG, automotive, healthcare, fitness, retail, legal, regulated CPG, and political verticals.
Built a CTV measurement framework for Thomas J. Henry that surfaced a 5-7 day reduced time-to-signup when CTV was layered with Display. Codebase handed off to Causal IQ at departure.
CTV measurement deep dive →
4 Python/Pandas tools I built on my own: bid optimization, audience insights, site list management, and campaign reporting. Automated workflows across TTD and DV360. Made the whole team 60x faster.
View Case Study →Managed $900K+/month in programmatic spend across 10+ DSPs (The Trade Desk, DV360, Yahoo, Zeta, and more), peaking at $1.2M/month during high-volume periods. Consistently ranked #1 or #2 out of 22 account managers. Contributed to a multimillion-dollar exclusivity deal with a key data partner.
AI fluency isn't a resume line - it's the operating layer. Three uses, each producing something concrete.
Claude Code, Codex, Cursor, and Copilot in rotation. Used to build the entire 950K-line JointCommerce platform and every project in the R&D lab.
Vercel AI SDK + Gemini powering the Stanford BIOE 230 contextual tutor. Real users, real prompts, real evaluation.
MCP-driven sales prospecting workflows at JointCommerce - automated account enrichment, outreach generation, and pipeline triage.
Verifiable adoption, codebase continuity, account growth, and team performance signals from the work itself - not paraphrased praise.
“Adopted Stephen’s BIOE 230 Interactive Platform for the Spring 2026 semester.”
“The CTV measurement framework Stephen built remained in use at Causal IQ after his departure, continuing to serve the broader client book.”
“Ranked #1 or #2 across the 22-person Account Operations team for performance discipline, optimization judgment, and client communication.”
“Made AI-assisted development the team-wide operating standard across engineering and operations - daily Claude Code, Codex, Cursor, and Copilot use by every team member. The team multiplier mattered more than any individual prompt skill.”
Direct recommendations from past managers, clients, and colleagues available upon request. LinkedIn recommendations at linkedin.com/in/stephenumunna.
Built a 100K-line learning platform in 24 hours.
Adopted by Stanford Bioengineering for Spring 2026.
Professor Todd P. Coleman · Stanford University · Department-wide rollout in progress
Technical proving grounds where I stress-test architecture patterns before applying them to production work. Healthcare data, finance, sports analytics, education - each one pushed a different skill that made the next company better.
100K-line learning platform adopted by Stanford Bioengineering. 180 interactive visualizations, contextual tutor, 960 practice problems.
Rates every US ZIP code 0-100 for healthcare access. Academic-grade spatial analysis, 4M+ providers, E2SFCA spatial scoring algorithm.
AI-powered personal finance for first-generation wealth builders. Link your bank, track spending, and see what your money could become if invested.
Full-stack NBA platform with live scores, player comparisons, shot charts, and 78 years of history. Built in 3 days. $0/month.
I'm exploring leadership roles where building the technology and running the business aren't treated as separate jobs.