About
Built by the person who needed it.
Quadfence is led by a practising dentist who made the decision this product exists to inform — and had almost nothing useful to make it with.
Founder
Dr. Soujanya Maddipati, DDS, MPH
Founder & Chief Executive
In December 2022 she bought a dental practice in North Houston. The seller's system showed patient counts, procedure mix and collections, and a CPA went through the books. Those numbers she could verify.
What nobody could tell her was anything about the market. The broker handed over generic census demographics — education, gender split, race. Not how many dentists per capita already competed for those patients. Not the renter and apartment ratio, which determines how fast a patient base turns over regardless of how well you run the practice. Not which clinical services the surrounding practices already offered, and which gaps were open.
She bought it anyway. She has run Mi Smile Family Dental ever since — Medicaid, CHIP, Medicare Advantage and PPO — and sees every patient herself. Then she spent three years buying marketing she could not measure, from agencies paid whether the practice grew or not.
She is the customer. That is the whole reason this exists.
Public health before dentistry. Population data analysis is the discipline this product automates — and the payer-mix complexity the engine has to model is the one she works inside every day.
The team
A physician director
Clinical and operational depth, and the bridge to the adjacent medical-practice market. Two clinicians who are the customer, in two neighbouring verticals.
A sales director
Owns onboarding and the agent motion across territories. Texas first, then parallel states — acquisition scales by adding agents, not overhead.
Chaitanya Maddipati
Co-founder & Chief Technology Officer
Twenty years building data and AI platforms inside regulated industries, twelve of them owning the product end to end — discovery, investment case, funding, delivery, adoption and measured benefit afterwards.
At Wells Fargo he was Executive Director and Principal Engineer over the bank’s 360-degree customer data platform: a forty-person organisation, a $15M multi-year portfolio, 3,000+ internal consumers. He ran the bank’s first application onto Google Cloud, negotiated the control model with Enterprise Architecture, Risk and Security, and wrote the reference architecture the rest of the bank built against. He retired five legacy stacks, moved 900TB onto BigQuery without a downstream consumer losing service, and delivered $2.5M of recurring annual savings validated after the fact against the original business case.
Before that, Senior Manager for Data and Cloud at Publicis Sapient, directing two global engineering pods across UnitedHealth Group and TikTok / ByteDance, where he architected a brand-measurement platform pre-win and led its delivery. Earlier, data products behind US credit-card acquisition at Capital One, where he led PCI-DSS tokenisation across nine warehouses, and enterprise modernisation at United Airlines.
He builds the engine underneath AiRadics: the consensus layer, the governance gate, the provenance log and the generation pipeline. The compliance-by-design approach is his — regulatory rules compiled into a gate that runs before generation rather than a review that happens after it.
Technical background
Regulated data platforms, at scale.
The problem AiRadics solves is not that generated marketing is hard to produce. It is that in a regulated category almost none of it can be published without being checked, and nobody can tell you afterwards why a model said what it said. That is the same problem he spent two decades on inside banks: decisions that have to be reproducible, auditable and defensible to a regulator.
Google Cloud is not a vendor choice made late. He ran a global bank’s first migration onto it and wrote the standards that followed.
Executive education in product management at Kellogg, digital marketing strategy at Harvard Business School and price optimisation at Wharton.
What Quadfence is
Governed AI for decisions that can be checked.
The engine underneath AiRadics is not specific to dentistry. Multiple AI models from competing vendors reason over the same question and have to agree before anything acts. Rules are compiled into code the engine is checked against. Every decision is logged so it can be reproduced.
Dental is where we prove it — a founder who is the customer, her own practice as the first place we try it, and a market fragmented enough to enter. Medical practices are the near-adjacent follow: the same ontology structure, different labels. Other local services come later, and not before dental and medical are both working.
Get in touch
If you run a practice, or you want the deck.
We are pre-product and looking for a small number of Texas practices to work with once there is something to work with. Investors: the deck and the operating model are available on request.