Indonesian · Singapore · Full Stack AI Engineer

I build AI platforms
from scratch.

I embed with teams to ship production AI systems — agents, pipelines, and the full-stack infrastructure behind them.

Edrick Kesuma
97%Faster AI agent build time2 weeks → 8 hours
85%AI bugs resolved autonomouslyAI-powered debugging pipeline
90%Cloud cost reductionServerless architecture refactor
5K+AI conversations served dailyWhatsApp CRM infrastructure
Top 1%Kaggle Computer Vision competitionPetfinder.my Pawpularity Contest
Hackathon ChampionSingapore

Client Work

AI Agents Built for Clients

Production agents deployed for real businesses — handling orders, locating showrooms, and automating workflows over WhatsApp.

Order Processing Agent

Reads parts lists from images, resolves stock conflicts, builds a cart, and generates a purchase order — end to end.

Showroom Locator Agent

Finds the nearest dealership from a shared location via WhatsApp and books a test drive in one conversation.

Internal Tools

AI Infrastructure Built for Companies

Production-grade internal platforms that let teams build, deploy, and maintain AI agents — without touching the underlying models.

Agent Builder
< 1 hrto deploy

A no-code wizard that lets any team build a fully configured production AI agent — from personality upload to live deployment.

Agent Debugger & Fixer
85%issues auto-resolved

Drop a failing trace into Slack. The system pulls logs, reads the agent's tools, diagnoses the root cause, and opens a fix PR — automatically.

About

From Dota to production AI

I got obsessed with AI in 2017 after watching OpenAI beat Dota 2 pros. Since then the path has run through a self-taught gap year, a Top 1% Kaggle finish in computer vision, 3x hackathon wins in Singapore, and eventually into production work at Covena building sales AI agents. I also share parts of that journey by making AI reels on Instagram.

Now I work as an AI Full Stack Engineer at Covena, where I build the agents, tooling, and infrastructure that help forward deployed engineers ship client AI agents faster and support more builds at the same time. That work has helped cut AI agent build time by 97%, resolve 85% of AI bugs autonomously, support 5K+ daily AI conversations, reduce cloud costs by 90%, and improve retrieval speed by 10x.

Agents and orchestration

I build end-to-end agent systems, from prompt and tool design to internal platforms that help teams create, test, and improve them faster.

Self-improving systems

I ship an MVP fast, learn from real usage, and build the eval and feedback loops that make the AI better over time.

Full-stack delivery

I ship the surrounding product too: Next.js and Express apps, APIs, CRM features, dashboards, WhatsApp integrations, Docker-based deployment, and observability.

LangChainLangGraphLangSmithQdrantNext.jsExpressDockerPostHogWhatsApp Cloud API

Current Role

AI Full Stack Engineer

Covena / VC-backed startup building sales AI agents

Now

I care most about shipping AI systems that help teams move faster, handle more client deployments, and stay maintainable after launch.

97%

faster AI agent build time

85%

AI bugs auto-resolved

90%

cloud cost reduction

5K+

daily AI conversations served

Core Areas

AI agentsRAGEvaluationsLLMOpsAI-based analyticsWhatsApp systemsFull-stack AI productsAI observability

Contact

If you're building with AI, let's talk.

Whether you're looking for a technical engineer who can actually ship, or you just want a second opinion on AI agents, evals, or architecture, I'd be happy to chat.

Email is the easiest way to reach me. If you want to browse a bit first, the rest of my socials are here too, along with my resume.

If you already have a rough use case, I can usually tell pretty quickly whether you need an agent, better tooling, or a simpler automation path.

Get In Touch

Best for hiring conversations, collaborations, or AI agent questions. Email is still the fastest route.