Terminal

chat · about
marckengland@brain: ~

$ whoami

Marck England · Full Stack Software Engineer · Generative AI Engineer

Manila · GMT+8

$ cat about.txt

I build software that solves real problems.

Shipped to production, used by real businesses every day. This whole page is one: ask me anything about the work.

* ask about my work, projects, or skills, and the graph lights up what I pull from.

>
Book a call Selected work Usually replies within 4 hours

The Brain

Knowledge graph · live
Waking the brain…
Active current
Me
Projects
Experience
Education
Skill areas
Technologies
Notes

132 nodes · 287 connections · click a node to trace it

Selected work

AI Investment Platform

FinTech · Private beta

An agentic advisor calls 15+ live tools and chains up to 5 rounds to answer in context, not guess.

A multi-asset investment platform with AI-assisted analysis, real-time market intelligence, and professional-grade risk controls. Built for retail investors who want institutional-quality tooling. Live in private beta with an invited user base.

  • Agentic AI advisor that calls 15+ live tools (your real portfolio, market data, correlations, dividends, forecasts, stress tests, and crisis scenarios), chaining up to 5 tool rounds to answer in context instead of guessing
  • AI-driven forecasting and buy/sell recommendations backed by risk scoring and geopolitical signals
Next.js 16TypeScriptPostgreSQLPrismaTanStack Query

Real-Time Fact Checker

Civic Tech · Private

Verdicts come back in seconds, checked against trusted Philippine fact-checkers.

A mobile app for iOS and Android. Share any link or screenshot from any app via the system share sheet. Within seconds, on-device classification triages the claim, and a backend NLP + RAG pipeline checks it against trusted Philippine fact-checking sources. Verdicts come back as Green (verified), Red (debunked), or Gray (needs review), and verified results auto-publish as articles anyone can browse.

  • Event-driven microservices: ingestion, NLP triage, verification, community, and notification services communicate over NATS JetStream, with pgvector RAG for retrieval
  • Share from any app: paste a Facebook post, drop a TikTok link, or hit Share in your browser
React NativeTypeScriptPythonFastAPIExpress

Autopilot Sales CRM

SaaS · Private beta

Build a follow-up cadence once and run it across thousands of leads on autopilot.

A multi-tenant sales CRM that automates the busywork of selling (activity logging, follow-up cadences, lead scoring, and forecasting), then surfaces each rep's single highest-value next action. Reps send and sync email from Gmail or Microsoft 365 / Outlook, call leads in-browser or via click-to-call, and migrate their book in from Salesforce or HubSpot in minutes.

  • Build a multi-step follow-up cadence once, in a step list or a drag-and-drop flow builder, and run it across thousands of leads on autopilot, with AI drafting a personalized email for every prospect
  • Built-in AI assistant answers questions and proposes next actions (log activity, draft a follow-up, update a deal) that reps approve in one click
Next.js 15TypeScriptSupabasePrisma 6OpenAI/Claude/DeepSeek + open-source LLMs (OpenRouter)

Voter Intelligence Platform

Civic Tech · In development

165,000+ bills tracked to their authors, and 77 controversies cross-referenced across 2+ outlets.

A platform that puts every Philippine politician's full track record in one place: bills they authored, controversies that follow them, family connections, and how they actually voted. Built on official records and fact-checked sources, so voters decide on data, not campaign jingles.

  • AI chatbot answers questions on any politician and cites official sources for every claim, provider-agnostic with admin-switchable models (Claude, Gemini, OpenAI, DeepSeek, or open-source models via OpenRouter)
  • AI research pipeline fills data gaps via grounded web search, accepting only facts confirmed by 2+ credible sources with full attribution, never unsourced
Next.js 16TypeScriptPostgreSQLPrismaElasticsearch

Sari-Sari Store POS

Retail · Desktop · Private

Keeps selling through brownouts and flags every peso that does not add up.

A point-of-sale system built specifically for Philippine sari-sari stores. Works through brownouts, turns any phone into a backup scanner, and is engineered to catch the cash that quietly disappears with discrepancy alerts, immutable logs, and daily snapshots that can't be edited away.

  • Optional AI assistant lets owners ask plain-language questions about sales and inventory, with strictly read-only access to store data (provider-agnostic: OpenAI, Anthropic, DeepSeek, OpenRouter, Gemini, or a custom endpoint)
  • Works through brownouts: keeps running offline, syncs everything the moment power returns
ElectronReactTypeScriptSQLitebetter-sqlite3

Numbers from shipped work, not adjectives.

24/7
autonomous support that hands off to your team only when you want it
RAG support agent, live since 2024
18
messaging apps & marketplaces unified into one inbox
WhatsApp, Messenger, Instagram, Shopee, TikTok, Shopify & more
165K+
congressional bills tracked to their authors
Voter Intelligence Platform

What I do.

Four things I build. Each one links to a real project that shipped and is running in production.

AI chatbots & agents

Assistants trained on your actual business (products, prices, policies) that answer accurately around the clock and hand off to your team the moment a human replies.

Resolves conversations end to end, escalating only the cases you choose to a human

Try the live terminal ↑

Custom web platforms

CRMs, dashboards, point-of-sale, and internal tools built around how your team actually works. One platform instead of five subscriptions stitched together.

Multi-tenant CRM running cadences across thousands of leads

Autopilot Sales CRM →

Workflow automation

Reminders that send themselves, payments that verify automatically, approvals reduced to one tap. Event-driven pipelines that free your people for real work.

Intake, classification, and reporting automated end to end

Real-Time Fact Checker →

Multi-channel messaging

Every customer message from every channel lands in one inbox, so your team answers from one place instead of juggling 18 apps.

18 channels in one inbox: WhatsApp, Messenger, Instagram, Shopee, TikTok, Shopify & more

How it was built →

Experience

Apr 2024 – Aug 2026

Full Stack Software Engineer · Generative AI Engineer

Intelligence Flow, Inc. (Tech Startup) · Taguig, Metro Manila
  • Founding Engineer, First Revenue-Generating Product: independently built the company’s first revenue-generating product during downtime, a 24/7 RAG customer-support agent that autonomously resolves conversations and escalates complaints, refunds, and complex cases to staff, with automatic handoff the moment a human agent replies (in production since 2024).
  • Omnichannel Messaging Platform: Unified 18 messaging apps and e-commerce marketplaces into a single inbox behind a pluggable per-channel integration layer with centralized webhook signature verification, so new channels could be added without changing the core.
  • Multi-Tenant SaaS Architecture: Re-architected the platform into a multi-tenant system with strict per-tenant data isolation, encrypted per-tenant credentials, and layered feature-flag and configuration controls.
  • Backend Re-Architecture (Flask → Litestar): Engineered the migration from Flask to an async Litestar + Piccolo ORM + Granian stack, structured with domain-driven design and a tenant-scoped repository pattern.
  • AI Orchestration & Cost Governance: Built agent orchestration on Pydantic AI across multiple LLM providers (Gemini, OpenAI, Anthropic, OpenRouter), with per-tenant token and cost tracking, configurable spend caps, and a per-tenant model allowlist.
  • Vector RAG Pipeline: Deployed Retrieval-Augmented Generation on PostgreSQL + pgvector with Gemini embeddings and background ingestion/embedding workers, delivering accurate, context-aware responses.
  • CRM Workflow Automation: Owned end-to-end delivery of a centralized CRM with AI-integrated messaging, replacing manual processes with one-click approval and denial workflows.
  • Vertical Product Build: Delivered an industry-specific vertical on top of the shared platform with profile management, an appointment scheduling calendar, and support ticketing.
  • Automated Appointment Reminders: Built a system that sends reminders 1–2 days before each appointment, improving scheduling reliability and reducing missed bookings.
  • AI Payment Verification: Delivered an AI service that scans receipt images for real-time payment validation and notifies the finance team for confirmation before a payment is marked as paid.
  • Secure Credit & Payments System: Designed a virtual-currency top-up system with voucher support and secure, scalable transaction handling.
  • Frontend Re-Architecture: Rebuilt the web client on React Router 7 (framework mode) + React 19 + Vite with TanStack Query, Jotai, Tailwind CSS v4, and i18next internationalization, following atomic design.
  • Generative AI Orchestration (n8n + LangChain): Built low-code n8n + LangChain workflows spanning Gemini and OpenAI for order intake, reply classification, and reporting.
  • Production Hardening & Observability: Instrumented the platform with Sentry, Apitally, and structlog, plus Redis-backed caching, rate limiting, and idempotency keys; containerized and deployed to Google Cloud Run.

Stack

The tools of the trade, grouped by where they live in a build.

Languages
PythonTypeScriptJavaScript
AI & LLM Engineering
LLMsVector RAGGraph RAGLangChainPydantic AIMCP (Model Context Protocol)Google GeminiOpenAIAnthropic ClaudeDeepSeekOpenRouterNLPn8n
APIs
Meta Graph APIMicrosoft GraphGoogle Workspace APILINE Messaging APITwilio (Voice)DiscordKindeResend
Cloud, DevOps & MLOps
DockerAWSGCPGoogle Cloud RunVercelRailwayRenderSupabaseDopplerInngestNATS JetStreamSentryCron JobsGit
Security
OAuth 2.0AES-256 encryptiontwo-factor authentication (2FA)Cloudflare Turnstileanti-crawler
Databases & Search
PostgreSQLMySQLMongoDBSQLitePrismaPiccolo ORMpgvectorElasticsearchRedis (Upstash)IndexedDB
Testing & Environments
JestVitestPlaywrightfast-checkHypothesis (property-based testing)Testing LibraryBiomePostmanJIRAAgile Scrum
Frameworks & Libraries
Next.jsReact.jsReact Router 7React NativeElectronLitestarNode.jsExpressFastAPIFlaskDjangoViteTanStack QueryJotaii18nextnext-intlTailwind CSSThree.jsshadcn/ui
Education

BS Computer Science

Pamantasan ng Lungsod ng Maynila · 2019–2023 · OGWA 1.71 · Dean's Lister (2020–2023)

ICT Strand

Metropolitan Medical Center - CAST · 2016–2018
Next project

Let's build something worth shipping.

Send a message

The more context you share, the more useful my first reply will be.

Every message lands in my personal inbox. I read them all.
Portrait of Marck England
Marck EnglandSoftware Engineer · Generative AI Engineer

Book a 30-min intro call

No pitch, no pressure. We'll talk through your project and whether I'm the right fit.

Book a call

Usually replies within 4 hours