I began my software development journey as a self-taught learner in middle school and pursued a technical programming course in high school. After high school, I worked as a full-stack web developer and then earned a Computer Science degree, gaining valuable experience at various companies, including NIC.br. There, I specialized in full-stack web and native Android development. Later, I co-founded Simpli, a startup that grew into a successful software house, delivering diverse projects, including blockchain development. This period helped me evolve as both a developer and a leader.

Two 2025 incidents, an agent that deleted a production database and one that nearly wiped a codebase, weren't prompt failures. They were boundary failures. A look at why AWS's security team, OWASP, and Okta are all converging on treating coding agents as a scoped, first-class identity, and why the same instinct has to extend to code review, not just access.

MCP officially deprecates Sampling on July 28 — the feature that let a third-party app borrow a user's own AI subscription instead of paying for inference itself. A look at which clients ever actually implemented it (Claude Desktop and Claude Code never did), why the "low adoption" rationale is circular, and the one pattern for free AI in your app that survives the change.

A developer's field guide to where AI and blockchain actually meet in 2026: agent payments with x402, on-chain agent identity with ERC-8004, AI in the contract writing and audit loop, and decentralized compute with verifiable inference. It separates what genuinely shipped from what's still a demo, and points to where a developer can start.

AWS Blocks is a new open-source TypeScript framework that turns your backend code into the AWS infrastructure that runs it, an approach called Infrastructure-from-Code. This piece covers what Blocks actually does, how it compares to earlier IFC tools like Encore.ts and Nitric, the lock-in and single-Lambda trade-offs worth watching, and why AWS's backing might finally push Infrastructure-from-Code into the mainstream.

Harness Engineering has two layers. The CI pipeline is one of them. This is the other — a structured discovery and spec workflow that constrains AI before it writes a single line of Solidity.
At 33Labs (formerly 33Audits), I've designed and implemented scalable backend services and application architecture for decentralized financial products.
The company began as a security auditing firm, which has given me daily experience preventing vulnerabilities in production systems. I've worked end to end — architecture, implementation, and automated testing — with security engineers involved throughout. This led me to a key conviction: review should start at the architecture stage, before any code is written. To support it, I built developer tooling and AI-assisted workflows that improved the team's productivity.
I co-founded Simpli as a two-person startup focused on building a B2C mobile product, which quickly evolved into a fast-growing software house serving a wide range of clients. In its first year, the company pivoted to delivering custom distributed applications and scaled organically through consistent delivery and client satisfaction. Over 11 years, we delivered 50+ successful digital products for both startups and enterprise clients — mostly React, Next.js and Node.js on the frontend and backend, plus Java/Kotlin services where a client's stack called for it.
I played a key role in shaping both the technical direction and business strategy of the company, leading system architecture, technical roadmaps, and documentation, and building an engineering org of 30 including 5 team leads. My work ranged from hands-on feature development to driving technology adoption and process design — helping turn product ideas into real businesses by aligning technical execution with market opportunities.
Provisioning AWS infrastructure was a recurring part of that work. With 50+ projects, we needed a standard: a checklist for spinning up a new account, a baseline set of IAM roles, and a reusable Terraform starter kit so no project's infra started from a blank page — the same template became a paved road that sped up every project's first deploy.
Enclave is a fintech-style application for secure digital asset management, built so non-technical users can onboard as smoothly as in a traditional web app. Working with a small team, I owned the product vision, usability, and the entire frontend — built in React and Next.js — contributed to the Node.js backend services, and led the system design of the indexer and API behind the app's activity explorer. One example of that indexing work: handling chain reorgs. Every block was stored keyed by its own hash and its parent's, so detecting a reorg meant walking back parent hashes until the indexer found one it already had — the fork point. From there, the orphaned blocks and the balances derived from them were invalidated, and the canonical chain was replayed forward through idempotent handlers, so reprocessing never double-counted a transaction.
Neon is the leading wallet in the Neo ecosystem, with over $1 billion in traded volume, built in React, React Native and Electron. I was responsible for architecting its mobile version and later contributed to the desktop app. During my time on the project, I tackled key challenges such as supporting multiple blockchain networks, managing multiple accounts simultaneously, implementing WalletConnect integration, and developing the protocol for network interaction, along with several other critical integrations.
Sharity was a crowdfunding platform for charitable causes that I built from creation through scaling, growing it past 100 thousand users. It was eventually sold to Abacashi, a bigger competitor, and I was invited to lead the engineering side of that merger in recognition of the quality of Sharity's system — refactoring Abacashi's legacy C#/Angular codebase over to Node.js and React with the platform staying live the whole time, using a Strangler Fig strategy to migrate it piece by piece instead of a big-bang rewrite. One example of the work on Sharity itself: an event-driven achievements system, with badges and real-time progress bars for challenges that could chain off one another. Business events flowed through SQS, and idempotency was enforced by inserting each event's ID into a dedupe table in the same MySQL transaction as the counter increment, so a duplicate delivery just aborted instead of double-counting; chained challenges unlocked themselves by publishing a follow-up event back onto the queue, with no need for global ordering. It held up at 100k+ users, and was later reused as-is inside Abacashi after the merger.
Wow Talents was a comprehensive agency platform connecting child models with modeling opportunities, supporting subscription payments and five distinct user types — models, guardians, agency staff, clients, and admins — each with its own view and permissions. I owned it end to end: turning a loosely defined set of demands into functional requirements and wireframes, designing the database architecture, and structuring the project from scratch. One example of that structuring work: since every model account belonged to a minor, every read and write had to be scoped through a guardian relationship as well as a role, so I designed the permission model around both the user type and that guardianship chain — letting new user types and rules get added later without touching the core matching logic.
Louis Dreyfus Company, one of the largest commodity traders in the world, commissioned a 'Safety, Health, and Environment' management platform for use across all its global units — each with its own local regulations, forms, and reporting requirements. I owned it end to end: architecture, development, and integration with Azure Active Directory for authentication and user management. The core challenge was translating that sprawl of per-unit requirements into one flexible application instead of one branch per country, so I modeled forms, fields, and workflows as configurable data instead of hardcoded screens — letting a compliance team in a new unit stand up its own SHE process by configuring the platform, not by requesting a code change and a deploy.
Jamef, the largest shipping company in Brazil, needed a new dashboard for customers to track delivery data, but the existing one suffered from severe performance problems. My scope started as frontend-only, but tracing the slowness back to its source pulled me into the data layer. I introduced a caching layer and a server-side pagination policy for the result set, taking a query that took 8 seconds down to a few milliseconds. I delivered a complex dashboard with several customized graphs and contributed to restructuring the underlying central system.
I was the lead engineer who built iTrack Brasil from the ground up, owning its system design through its first two years and scaling it into a B2B delivery platform integrating multiple systems, with nearly 60,000 couriers and over 50 million invoices processed across 2,000 registered companies — growth that led to its acquisition by MadeiraMadeira in 2021. Simpli carried the product through that acquisition and ran it for three more years before handing it over to MadeiraMadeira in 2024, with me on it as CTO, coming back alongside the team hands-on when the work needed the depth. One of those returns: as invoice volume and courier position updates grew, the database hit a write bottleneck, surfaced by a spike at the end of a fiscal month. The system stayed up, but latency and AWS costs climbed, so I redesigned the write path at its source — courier position pings now land as Parquet files on S3, backed by a Redis cache holding each courier's last known position for fast reads, while invoice processing moved behind an SNS/SQS pipeline that absorbs bursts asynchronously.
Mapix connects students with their school drivers, giving children and parents predictability and safety while simplifying route planning and communication for drivers. I owned the planning and structuring of the solution end to end and built its most critical components: the route-building system, in-app chat, and a live GPS mode. One example of that work: building a route wasn't just point-to-point — each van carried several students with different pickup and drop-off points on a shared trip, so the routing logic had to sequence stops for the whole group at once, and the GPS mode had to keep parents' ETAs accurate against a background location stream over unreliable mobile connections, without draining a driver's phone battery on a multi-hour shift.
Apptite was a food delivery app for iOS, Android and the web. It gained recognition with acceleration by '500 Startups' and media coverage that established it as an important platform in the artisanal food market. I was the main engineer responsible for the platform from its initial planning and structuring through scaling it as it grew. One example of that scaling work: the dish recommendation engine. The map was discretized into cells, each holding the list of stores that served it, updated in batch whenever delivery areas changed — so the geographic lookup itself was just a cached read. Personalization ran on top of that already-filtered set, combining complementary signals in parallel (repurchase history, regional popularity, among others), with the heavier estimates — including a store-by-cell matrix — precomputed offline.
Desabafa was an anonymous social network built for emotional support and mutual understanding, and it went on to receive media recognition in the mental health sector while reaching 700 thousand posts. I was responsible for planning, structuring, and building the platform, including its moderation and safety systems. One example of that work: anonymity was the whole point of the product, so any moderation approach that relied on identifying users to punish or filter them was off the table. I built the safety layer around behavior instead of identity — rate-limiting and pattern detection on posting activity, plus keyword and pattern-based flagging tuned for crisis language — so harmful content and abuse could be caught and escalated without ever unmasking a user.
Bandeirantes, one of Brazil's largest media conglomerates, has run Band Radios since 1937 and selected my team in 2014 to modernize its mobile app. I served as lead developer for both the Android and iOS apps. One example of the technical work: streaming ran over UDP, and the naive approach — opening the connection only after a listener picked a station — meant a noticeable delay before audio started. I pre-established the UDP connection ahead of the station choice instead, trading a bit of idle network and battery use for audio that started the instant a user tapped a station.
Multilaser, one of Brazil's largest cell phone and tablet manufacturers, faced high demand for quality control testing that had been done by hand on the assembly line. I led development of the Android app that automated it instead — running CPU, RAM, GPS, screen brightness, and touch checks — and it has since tested over 20 million devices. One example of that work: each check had its own pass/fail criteria and its own margin for hardware variance, so the real problem wasn't running the tests, it was collapsing five independent, noisy hardware signals into one deterministic pass/fail a line worker could act on in a couple of seconds, fast enough to keep up with a production line moving tens of thousands of units.
At NIC.br, I worked on SIMET, Brazil's official internet quality measurement tool used by regulators and ISPs alike. The flagship SIMET application still ran as a Java Applet at a time when browsers were starting to lock those out for security reasons, so I proposed and led its migration to plain JavaScript before that became an emergency. Beyond the core tool, I built SimetMapas, visualizing internet quality heat maps across Brazil, and dashboards consumed directly by internet operators and regulatory agencies. I also contributed to SimetBox, a Wi-Fi router built to run these tests automatically, and to an Android testing app with its own custom graphics library for rendering results on constrained hardware.
My first three engineering jobs — web applications in PHP and Java, JavaScript front-ends and MySQL.
At 33Labs (formerly 33Audits), I've designed and implemented scalable backend services and application architecture for decentralized financial products.
The company began as a security auditing firm, which has given me daily experience preventing vulnerabilities in production systems. I've worked end to end — architecture, implementation, and automated testing — with security engineers involved throughout. This led me to a key conviction: review should start at the architecture stage, before any code is written. To support it, I built developer tooling and AI-assisted workflows that improved the team's productivity.
I co-founded Simpli as a two-person startup focused on building a B2C mobile product, which quickly evolved into a fast-growing software house serving a wide range of clients. In its first year, the company pivoted to delivering custom distributed applications and scaled organically through consistent delivery and client satisfaction. Over 11 years, we delivered 50+ successful digital products for both startups and enterprise clients — mostly React, Next.js and Node.js on the frontend and backend, plus Java/Kotlin services where a client's stack called for it.
I played a key role in shaping both the technical direction and business strategy of the company, leading system architecture, technical roadmaps, and documentation, and building an engineering org of 30 including 5 team leads. My work ranged from hands-on feature development to driving technology adoption and process design — helping turn product ideas into real businesses by aligning technical execution with market opportunities.
Provisioning AWS infrastructure was a recurring part of that work. With 50+ projects, we needed a standard: a checklist for spinning up a new account, a baseline set of IAM roles, and a reusable Terraform starter kit so no project's infra started from a blank page — the same template became a paved road that sped up every project's first deploy.
Enclave is a fintech-style application for secure digital asset management, built so non-technical users can onboard as smoothly as in a traditional web app. Working with a small team, I owned the product vision, usability, and the entire frontend — built in React and Next.js — contributed to the Node.js backend services, and led the system design of the indexer and API behind the app's activity explorer. One example of that indexing work: handling chain reorgs. Every block was stored keyed by its own hash and its parent's, so detecting a reorg meant walking back parent hashes until the indexer found one it already had — the fork point. From there, the orphaned blocks and the balances derived from them were invalidated, and the canonical chain was replayed forward through idempotent handlers, so reprocessing never double-counted a transaction.
Neon is the leading wallet in the Neo ecosystem, with over $1 billion in traded volume, built in React, React Native and Electron. I was responsible for architecting its mobile version and later contributed to the desktop app. During my time on the project, I tackled key challenges such as supporting multiple blockchain networks, managing multiple accounts simultaneously, implementing WalletConnect integration, and developing the protocol for network interaction, along with several other critical integrations.
Sharity was a crowdfunding platform for charitable causes that I built from creation through scaling, growing it past 100 thousand users. It was eventually sold to Abacashi, a bigger competitor, and I was invited to lead the engineering side of that merger in recognition of the quality of Sharity's system — refactoring Abacashi's legacy C#/Angular codebase over to Node.js and React with the platform staying live the whole time, using a Strangler Fig strategy to migrate it piece by piece instead of a big-bang rewrite. One example of the work on Sharity itself: an event-driven achievements system, with badges and real-time progress bars for challenges that could chain off one another. Business events flowed through SQS, and idempotency was enforced by inserting each event's ID into a dedupe table in the same MySQL transaction as the counter increment, so a duplicate delivery just aborted instead of double-counting; chained challenges unlocked themselves by publishing a follow-up event back onto the queue, with no need for global ordering. It held up at 100k+ users, and was later reused as-is inside Abacashi after the merger.
Louis Dreyfus Company, one of the largest commodity traders in the world, commissioned a 'Safety, Health, and Environment' management platform for use across all its global units — each with its own local regulations, forms, and reporting requirements. I owned it end to end: architecture, development, and integration with Azure Active Directory for authentication and user management. The core challenge was translating that sprawl of per-unit requirements into one flexible application instead of one branch per country, so I modeled forms, fields, and workflows as configurable data instead of hardcoded screens — letting a compliance team in a new unit stand up its own SHE process by configuring the platform, not by requesting a code change and a deploy.
Jamef, the largest shipping company in Brazil, needed a new dashboard for customers to track delivery data, but the existing one suffered from severe performance problems. My scope started as frontend-only, but tracing the slowness back to its source pulled me into the data layer. I introduced a caching layer and a server-side pagination policy for the result set, taking a query that took 8 seconds down to a few milliseconds. I delivered a complex dashboard with several customized graphs and contributed to restructuring the underlying central system.
I was the lead engineer who built iTrack Brasil from the ground up, owning its system design through its first two years and scaling it into a B2B delivery platform integrating multiple systems, with nearly 60,000 couriers and over 50 million invoices processed across 2,000 registered companies — growth that led to its acquisition by MadeiraMadeira in 2021. Simpli carried the product through that acquisition and ran it for three more years before handing it over to MadeiraMadeira in 2024, with me on it as CTO, coming back alongside the team hands-on when the work needed the depth. One of those returns: as invoice volume and courier position updates grew, the database hit a write bottleneck, surfaced by a spike at the end of a fiscal month. The system stayed up, but latency and AWS costs climbed, so I redesigned the write path at its source — courier position pings now land as Parquet files on S3, backed by a Redis cache holding each courier's last known position for fast reads, while invoice processing moved behind an SNS/SQS pipeline that absorbs bursts asynchronously.
Apptite was a food delivery app for iOS, Android and the web. It gained recognition with acceleration by '500 Startups' and media coverage that established it as an important platform in the artisanal food market. I was the main engineer responsible for the platform from its initial planning and structuring through scaling it as it grew. One example of that scaling work: the dish recommendation engine. The map was discretized into cells, each holding the list of stores that served it, updated in batch whenever delivery areas changed — so the geographic lookup itself was just a cached read. Personalization ran on top of that already-filtered set, combining complementary signals in parallel (repurchase history, regional popularity, among others), with the heavier estimates — including a store-by-cell matrix — precomputed offline.
At NIC.br, I worked on SIMET, Brazil's official internet quality measurement tool used by regulators and ISPs alike. The flagship SIMET application still ran as a Java Applet at a time when browsers were starting to lock those out for security reasons, so I proposed and led its migration to plain JavaScript before that became an emergency. Beyond the core tool, I built SimetMapas, visualizing internet quality heat maps across Brazil, and dashboards consumed directly by internet operators and regulatory agencies. I also contributed to SimetBox, a Wi-Fi router built to run these tests automatically, and to an Android testing app with its own custom graphics library for rendering results on constrained hardware.
My first three engineering jobs — web applications in PHP and Java, JavaScript front-ends and MySQL.
A selection of projects — the complete history, including those under Notable Achievements, is at https://gil.solutions
Bachelor's degree in Computer Science from Pontifícia Universidade Católica de São Paulo, one of Brazil's leading higher education institutions.

Two 2025 incidents, an agent that deleted a production database and one that nearly wiped a codebase, weren't prompt failures. They were boundary failures. A look at why AWS's security team, OWASP, and Okta are all converging on treating coding agents as a scoped, first-class identity, and why the same instinct has to extend to code review, not just access.

MCP officially deprecates Sampling on July 28 — the feature that let a third-party app borrow a user's own AI subscription instead of paying for inference itself. A look at which clients ever actually implemented it (Claude Desktop and Claude Code never did), why the "low adoption" rationale is circular, and the one pattern for free AI in your app that survives the change.

A developer's field guide to where AI and blockchain actually meet in 2026: agent payments with x402, on-chain agent identity with ERC-8004, AI in the contract writing and audit loop, and decentralized compute with verifiable inference. It separates what genuinely shipped from what's still a demo, and points to where a developer can start.

AWS Blocks is a new open-source TypeScript framework that turns your backend code into the AWS infrastructure that runs it, an approach called Infrastructure-from-Code. This piece covers what Blocks actually does, how it compares to earlier IFC tools like Encore.ts and Nitric, the lock-in and single-Lambda trade-offs worth watching, and why AWS's backing might finally push Infrastructure-from-Code into the mainstream.

Harness Engineering has two layers. The CI pipeline is one of them. This is the other — a structured discovery and spec workflow that constrains AI before it writes a single line of Solidity.