César Rodríguez
16 years in finance. Now I build the systems.
CFO of a US biotech, compliance leader at CNBV-regulated institutions, and winner of Mexico's Fintech Sandbox Challenge. Today I design, fund, and ship AI-native financial platforms — compliance engines, trading systems, investor portals — with the architectural depth regulated markets demand.
Three archetypes. One operator.
Regulated products fail in the gaps between the lawyer, the CFO, and the engineering team. I operate where the three overlap — the regulation defines the spec, the capital defines the constraints, and the code ships with both built in.
Projects & Platforms
Platforms across compliance, trading, business, and education verticals. Built with modern AI-assisted development — every metric shown maps to a verifiable fact in the codebase.
SENTINEL PLD/AML
Local-AI PLD/AML pipeline for CNBV-regulated institutions
End-to-end PLD/AML document pipeline — four-engine OCR cascade, local SLM extraction, sanction-list screening with Spanish phonetic matching, and human-in-the-loop review — with AI running 100% locally so client documents never leave the deployment.
ESCUDO PLD/FT
PLD/FT risk platform for LFPIORPI-regulated entities
Multi-tenant PLD/FT platform for entities under Mexico's LFPIORPI: explainable IRI/IRR risk methodology, list screening, ten alert typologies, SPPLD notice drafting, and Excel bulk onboarding — in active development toward first staging.
Identia
Privacy-first compliance OS for Mexico's PUI regulated entities
Multi-tenant compliance SaaS for PUI-regulated financial institutions that runs CURP screening entirely hash-vs-hash — personal data is never stored or transmitted in clear — plus the full governance workload around it.
KYC Flow
On-prem DDQ/KYC copilot that cites evidence or refuses to answer
On-premise copilot that helps compliance officers answer KYC/DDQ questionnaires with literally cited evidence from their own document vault, running LLMs 100% locally so confidential documents never leave the machine.
CR Engine
Stock trading analysis with insider tracking and scoring
Multi-factor stock analysis engine that crosses quantitative scoring with real-actor intelligence — SEC Form 4 insider tracking — and explains every signal it produces.
Options Engine
Crypto options trading with automated strategy execution
BTC options trading engine with Deribit integration, straddle strategies, and automated trigger-based execution.
CryptoExchange
Full-featured BTC/USDT exchange with order book
Complete cryptocurrency exchange platform with real-time order book, market/limit orders, and CoinGecko price integration.
PropTech CDMX
Real estate opportunity analysis for Mexico City
Data-driven real estate platform analyzing purchase opportunities in CDMX with profitability metrics and zone comparisons.
Investor Portal
White-glove investor portal for private real-estate sponsors
Premium LP-facing portal for private real-estate projects: position, XIRR/MOIC/DPI performance, distributions, secure data room, and capital account — with investor isolation enforced at the database layer.
WhatsApp InfoBot
Replicable WhatsApp assistant with zero invented answers by design
WhatsApp assistant for sales, customer support, and service desk that answers only from each business's curated knowledge base — and escalates to a human when it doesn't know.
Club Glow
Loyalty platform where the balance is replayed, never stored
Points-based loyalty program running in production for a Mexican beauty business, where a customer's balance is never stored — it is replayed from an append-only ledger, so the number on screen cannot drift from the transactions behind it.
How I Build
Most builders read the regulation, model the P&L, or write the code. Regulated markets need all three — the path from a disposición to a production URL an auditor can trust.
const edge = "I understand the regulation, the capital, AND the code";
- 01Business
Regulation & business model
I start from the disposición and the P&L, not the ticket — in regulated markets the law is the spec.
Obligations from LFPIORPI, the PUI, CNBV and LFPDPPP mapped straight to data models and controls; every field classified by sensitivity before a schema exists.
Regulatory mappingData classificationUnit economics - 02Architecture
Architecture & trust boundaries
Where the sensitive data lives is what lets a regulated buyer say yes — I design that boundary first.
Deny-by-default RLS with a cross-tenant leak test in CI, hash-vs-hash so PII never travels in clear, a documented STRIDE model, and one portable Compose core for cloud, VM or on-prem.
PostgreSQL RLShash-vs-hashSTRIDEDocker Compose - 03Engineering
Test-first build
Business logic that's wrong costs a fine, not a bug report — so the rules engine gets a failing test before it gets code.
TDD on validators, RBAC and rules engines; AI-assisted execution under human-directed architecture; local-first LLMs (vLLM / Ollama) with pgvector, so documents never leave the machine and answers cite evidence or refuse.
TypeScriptPython / FastAPIpgvectorvLLM / Ollama - 04Security
Security & auditability
An auditor should be able to verify the system without trusting me — that verifiability is a feature, not a document.
Append-only hash-chained audit (triggers block UPDATE / DELETE), HMAC-signed exports, TOTP MFA, TLS 1.3 and encryption at rest, and backup / DR drills with measured RPO / RTO.
hash-chain auditTOTP MFAHMAC exportsTLS 1.3DR drills - 05Delivery
Ship & prove
I hand over a working URL and an evidence file the compliance officer can take to their board — piloted on real data, not a demo seed.
CI/CD gated by the cross-tenant test, containerized deploys to Vercel, Railway or a dedicated VM, runbooks and a regulator-ready evidence kit.
GitHub Actions CIVercelRailwayevidence kit
- Frontend
- Next.jsReactTypeScriptTailwindRecharts
- Backend
- PythonFastAPINode.jsDeno edgeInngest
- Data
- PostgreSQLSupabasepgvectorRedisS3 / MinIO
- AI · local-first
- vLLMOllamaClaudepgvector RAG
- Security
- RLS multi-tenanthash-chain auditTOTP MFATLS 1.3SSE-S3
- Delivery
- DockerVercelRailwayGitHub ActionsDR drills
How I Work With You
Compliance & RegTech Systems
PLD/AML pipelines, KYC/DDQ automation, and audit-ready systems for regulated institutions — designed to run inside the client's own infrastructure. Delivered through Gravisphere Intelligence, my boutique consultancy.
Visit GravisphereTechnology Consulting
AI-driven process automation and digital adoption for businesses ready to modernize: AI agents and workflow orchestration, systems integration, data pipelines, and custom internal tools — from opportunity mapping to deployed, measured automation.
Visit GI ConsultingAngel Investment
Early-stage capital for AI, fintech, and digital products. I evaluate founder-market fit, capital efficiency, and fast validation cycles.
GitHub
Select open-source projects and code samples. Most production systems remain private.
BTC options trading engine with Deribit API integration and automated strategy execution
Let's Build Something
Looking for an angel investor, technical advisor, or execution partner who understands both the regulation and the code? Let's talk.