Open to new rolesSingapore

Hey, I'm Ismail. I build AI systems that make it to production.

Ismail Jaija

I embed with a team, find the real problem, and ship the thing that solves it in their environment, not a demo sandbox. I was first engineer on an enterprise security platform that now runs inside fifteen banks and insurers.

6+
Years shipping software
20+
Projects delivered
0→1
MVP to enterprise production
15
Enterprise deployments

Services

How I work alongside your team

Embedded engagements rather than handovers: your stakeholders, your stack, your constraints — from the first discovery call to the thing running in production.

Discovery & prototyping

01

Sitting with the people who will actually use it to work out what the product has to do — then putting something working in front of them in weeks, so we find out I was wrong while it is still cheap.

  • Stakeholder discovery
  • Requirements
  • Rapid prototyping
  • Technical demos
  • Scoping

Agentic AI that holds up

02

LLM agents, RAG and MCP servers wired into your own services so they can do work rather than describe it — with schemas, guardrails and evaluation, because a demo that impresses is not a system you can run.

  • Claude Agent SDK
  • MCP
  • RAG
  • Evaluation pipelines
  • Tool use
  • Vector search

Backends & real-time systems

03

Event-driven services and real-time pipelines behind clean REST and gRPC APIs — the part that has to keep working at three in the morning, instrumented for production from the first commit.

  • Go
  • Python
  • FastAPI
  • Node.js
  • gRPC
  • PostgreSQL
  • Redis

Deployment & integration

04

Getting it into your environment and keeping it there: CI/CD, containers, infrastructure-as-code and observability — plus being the person who can drop to the layer where it is actually broken when something fails on your side.

  • Kubernetes
  • Docker
  • Terraform
  • GitLab CI
  • Prometheus
  • AWS · GCP · Azure

About

Close to the customer, deep in the stack

I have been building for other people's businesses since 2017 — scoping directly with stakeholders across Europe, the USA and the Gulf, then shipping what they turned out to need rather than what the brief said on day one. Alongside that I went the other direction, down into systems fundamentals — the layer under the frameworks, in C and C++.

Sekera Group is how I broke into startups. I joined as its founding engineer intern and left as Founding Engineer / Team Lead, taking the SIEM platform from MVP to production inside fifteen banks and insurers — regulated environments where shipping means meeting the customer's constraints, not just passing your own tests — and growing the team from one engineer to eight.

That combination is the point. The hard part of deploying into someone else's environment is rarely the code: it is the discovery, the constraints nobody wrote down, and being able to drop to the layer where a thing is actually broken while the customer is watching. I prototype fast and make it provable — then work security-first, with code review, TDD and design patterns as defaults, not afterthoughts.

Toolkit

Languages
PythonGoTypeScriptC / C++
Backend & AI
FastAPINode.jsClaude Agent SDKMCPRAGEvaluation pipelines
Frontend
ReactNext.js
Databases
PostgreSQLMongoDBOpenSearchClickHouse
Cloud & DevOps
AWSDockerKubernetesGitLab CITerraform
Security & Architecture
OWASP Top 10MicroservicesEvent-drivengRPCService Mesh

Based in Singapore. Working in English, French, and Arabic.

Experience

Where I've been building

  1. Education

    Sep 2026 — Present

    Current

    Cloud & AI Architecture

    SUTD · 42Singapore

    Singapore University of Technology and Design.

    • Cloud and AI architecture, continuing the systems-first training that started at 42Morocco.
  2. Role

    Jul 2025 — Sep 2026

    Founding Engineer / Team Lead

    Sekera Group

    A cybersecurity software startup.

    • Took the SIEM platform from MVP to production, now serving 15 established banks and insurance companies, owning architecture and stack decisions.
    • Designed and shipped agentic AI workflows that automate analyst triage and investigation tasks.
    • Grew engineering from 1 to 8 engineers; led Agile/Scrum, mentored contributors, and ran CI/CD and production releases on GitLab CI.
  3. Role

    Mar 2025 — Jul 2025

    Founding Engineer (Intern)

    Sekera Group

    First engineer on the product.

    • Built the initial SIEM MVP: backend services, core architecture, REST APIs, and real-time data pipelines.
    • Shipped the v1 that landed the first enterprise customers.
    • Established automated testing and the first CI/CD pipeline.
  4. Role

    Mar 2023 — Feb 2025

    Full-Stack Software Engineer

    Freelance

    End-to-end delivery for e-commerce and entertainment clients.

    • Delivered end-to-end software for clients across Europe, the USA and the Gulf, in Go, Python, Node.js, React and Vue.js.
    • Scoped requirements directly with client stakeholders and shipped scalable backends and REST APIs on AWS and Azure.
    • Set up automated testing, CI/CD, and monitoring for high availability and rapid iteration.
  5. Education

    Oct 2023 — Mar 2025

    Digital Technologies Architecture

    UM6P · 42Morocco

    Université Mohammed VI Polytechnique.

    • Built low-level systems in C and C++98: an HTTP/1.1 server, concurrency primitives, and network programming.
    • Practiced Agile with weekly sprints, code reviews, and pair programming.

Projects

Selected work

A few things I've designed and shipped — spanning agentic AI, cloud platforms, GitOps delivery, and market data.

Agentic forecasting SaaS01

SKUSeer

A platform that auto-selects and switches between forecasting models based on each SKU's data characteristics.

  • Ingests external feeds — weather, economic indicators, news sentiment — to learn which factors are truly predictive per product.
  • Runs continuous A/B testing across model combinations, adapting per category: seasonality-tuned for fashion, weather-adjusted for grocery.
  • Shipped an MVP, iterated to a POC, and secured commitments from early adopters across retail and logistics.
  • Python
  • FastAPI
  • React
  • ARIMA
  • Prophet
  • LSTM
  • XGBoost
  • PostgreSQL
  • Docker
MCP server for AI assistants02

Casablanca-Bourse Lens

Gives AI assistants live access to the Casablanca Stock Exchange — quotes, fundamentals, macro, FX and corporate actions — through plain-language questions.

  • A tool surface covering market overviews and stock screening, company fundamentals and shareholders, Bank Al-Maghrib macro data, FX and commodity prices, news sentiment, and corporate actions.
  • Every tool talks to a provider interface, so sources — WafaBourse, Bank Al-Maghrib, Yahoo Finance, AMMC filings — can be swapped without touching a single tool definition.
  • A DuckDB cache with per-dataset freshness — seconds for quotes, hours for macro, a day for company profiles — keeps answers fast without hammering upstream sources.
  • Built for the BVC annual trading competition, and used from Claude Desktop.

~70% less research time · ~35% better portfolio return vs. prior year

  • Python
  • MCP SDK
  • DuckDB
  • SQLite
  • Claude Desktop
Real-time multiplayer · AI-native03

Agentic Gaming Platform

An agentic gaming and social platform combining real-time multiplayer gameplay, live communication, and AI-native user experiences.

  • Scalable React and Django architecture with LLM-powered assistants and autonomous agents that interact directly with application services rather than acting as standalone chatbots.
  • WebSocket event streams carrying gameplay, chat and presence over one connection, served by the same ASGI application as the REST API.
  • Secure authentication across JWT, OAuth, Google SSO and TOTP two-factor, plus rankings, matchmaking and social features.
  • Django
  • Django REST Framework
  • Channels
  • React
  • WebSockets
  • JWT
  • OAuth
  • PyOTP
Multi-service orchestration04

Cloud Infrastructure Platform

A multi-service cloud architecture with automated orchestration and full observability on GCP.

  • Docker and Kubernetes with automated orchestration and health-checked, rollback-capable deployments.
  • Prometheus and Grafana monitoring stack for real-time observability and alerting.
  • Nginx reverse proxy with SSL/TLS termination, load balancing, and security headers.
  • Docker
  • Kubernetes
  • Prometheus
  • Grafana
  • Nginx
  • GCP
Agentic SRE copilot05

OpsPilot

An AI agent that investigates production alerts end to end — correlating metrics, logs, and recent deploys to surface root cause and a proposed fix.

  • Ingests firing Prometheus / Alertmanager alerts, then runs an agentic loop over live cluster state — metrics, pod logs, events, and recent GitLab deploys — to isolate root cause.
  • Drafts a remediation runbook and opens a GitLab merge request or proposes a kubectl action, gated behind human approval and read-only RBAC by default.
  • Grounds its answers with retrieval (RAG) over past incidents and runbooks, so recommendations reflect how this team has actually resolved similar issues before.
  • Exposed as an MCP server so on-call engineers can query cluster health and incident history in natural language, with findings posted to Slack.
  • Go
  • Kubernetes
  • Prometheus
  • LangChain
  • RAG
  • MCP SDK
  • GitLab CI
  • Slack
Kubernetes · Argo CD06

GitOps Delivery Platform

A GitOps workflow where the cluster's desired state lives in Git and is continuously reconciled — every change auditable and reversible.

  • Provisioned lightweight K3s / K3d clusters as reproducible infrastructure-as-code, standing up a full environment from scratch in a single command.
  • Built continuous delivery with Argo CD: application manifests declared in Git and automatically synced to the cluster on every commit.
  • Configured ingress routing and namespace isolation across multiple apps, with instant rollbacks by reverting a commit.
  • K3s
  • K3d
  • Argo CD
  • Kubernetes
  • Vagrant
  • GitLab CI
IaC · Ansible07

Automated Cloud Deployment

One-command provisioning of a resilient, reproducible web stack on a cloud server — from a bare VM to a running application.

Built with Bilal Eddinaoui

  • Automated the full server bootstrap with Ansible: packages, hardening, Docker, and the application stack in a single idempotent playbook.
  • Containerized a multi-service stack with Docker Compose behind an Nginx reverse proxy with TLS termination.
  • Engineered for resilience and reproducibility — persistent volumes, scheduled backups, and a full rebuild on a fresh server straight from code.
  • Ansible
  • Docker
  • Docker Compose
  • Nginx
  • Cloud VPS

Recognition

Distinctions

Hackathon wins, a consulting case podium, and a published handbook.

  • McKinsey Tech2Impact Hackathon

    Finalist — AI-powered sustainability solution

  • UM6P × Human Collective Intelligence Hackathon

    1st place — innovative AI solution

  • Cursor × AI Crafters Hackathon

    3rd place

  • PwC Consulting Case Competition

    3rd place — business strategy

  • MVP Essentials Guide (2025)

    Authored a handbook for IT startup founders

Contact

Have something to build? Let's talk.

Tell me about your project — infrastructure, backend, AI, or a full MVP. I'll get back to you within a couple of days.

Chat on WhatsApp

The fastest way to reach me — I usually reply within a few hours.

Book a 30-min intro call

or fill out the form