Open to new opportunities

Fernando Ott

Head of AI & AI Architect at 8 Figure Agency

I design multi-agent AI systems that plug into real business tools, understand context, and make decisions autonomously. My thing is turning complex AI ideas into platforms that actually run in production and move the needle for clients. 16+ years shipping software, 4+ years deploying AI at scale.

LinkedInGitHubYouTube8FAI Profile ↗
Fernando Ott

Key Metrics

16+
Years Engineering
4+
Years AI in Production
98
People Led in Career
25M+
Users Reached in Prod
2
New Lives in the World
What I Bring

Where I'm strongest

These aren't buzzwords. They're the things I've done consistently across startups, enterprise, and AI systems for over a decade.

Expert🏗️

AI Systems Architecture

I design multi-agent platforms from scratch. Brain is a multi-tenant OS with vector, graph, and structured stores, LLM orchestration, ingestion pipelines, and production observability. I think in platforms, not scripts.

Expert🌉

Business ↔ Engineering Bridge

I translate what a company needs into what the engineering team builds. This is the "click" that most developers miss. I've done it as CTO, Tech Lead at a NASDAQ company, and Head of AI at an agency.

Expert👥

People Management

I've managed nearly 20 people directly. Teaching juniors, managing senior egos, keeping motivation high. The code is simpler than people, but it's really satisfying when you figure out your way to lead and the team appreciates it.

Expert🤝

Client Conversation & Management

From agencies to NASDAQ enterprises, I sit with clients, understand their context, map AI to their workflows, and deliver measurable outcomes. Not demos, not slide decks. Real results they can see in their numbers.

Expert🔧

LLM Orchestration & Agentic AI

LangChain, CrewAI, MCP Server, multi-model pipelines, RAG, tool use, guardrails. I define prompting patterns, retrieval logic, evaluation frameworks, and failure modes. This is what I do daily in production.

Strong🚀

POC → Production Delivery

Brain went from whiteboard to multi-tenant production in months. KeHE search went from broken to 14x faster. Polen went from idea to 25M users. I take things from zero to one and then from one to scale.

Expert📐

Scalable Platform Thinking

I don't build one-off automations. I build platforms. Multi-tenant isolation, reusable agent patterns, metadata-driven configuration. Adding a new client to Brain doesn't require custom engineering, just configuration.

Strong☁️

AWS Production Workloads

At KeHE, AWS was the backbone: Lambdas, SQS, S3, DynamoDB, OpenSearch, CodeBuild, CloudWatch. I've designed and deployed AI/ML workloads end-to-end on AWS. No certifications yet, but deep hands-on experience.

Strong🎯

Startup Founder Mindset

I've co-founded, raised VC ($550K), served as CEO and CTO, lived through accelerators in 3 countries. I know the rhythm of fast delivery and user-first thinking. I can balance business urgency with technical integrity.

Where I'm honest about gaps

No one is strong at everything. Here's where I'm growing and what I consciously delegate.

AI Image & Video GenerationI understand the concepts and can architect integrations, but I haven't built production pipelines with diffusion models, Stable Diffusion, or video generation tools. I'd delegate this to a specialist.
Classical ML / Deep Learning ResearchMy ML work is applied and production-focused: RAG, LLM orchestration, fine-tuning for domain adaptation, multi-model routing. I don't train CNNs from scratch or write custom loss functions.
Mobile DevelopmentI've built web and backend systems for 16+ years, but no native iOS or Android experience. For products that need native mobile clients, I'd bring in a specialist.
PowerBII've queried ClickHouse and BigQuery in production contexts, but BI tooling — dashboards, data modelling, semantic layers — has never been my focus. Not where I want to specialize.
▶ YouTube · Otimiza AI

Sharing what I know about AI

Practical content on building AI systems, agents, and automation — in Portuguese. Subscribe →

EN

Vibe Coding Levels 1–5: Most Devs Are Stuck at 3

A breakdown of the 5 levels of vibe coding — from manual copy-paste to fully autonomous agent teams. Includes a live demo of my /launch skill spawning parallel agents to build 17 integrations end-to-end.

PT-BR

Create & Edit Videos Fast with These 3 AI Tools (Hack #03)

How I edit long videos in 20 minutes using Tella, CapCut, and Captions — automatic cuts, shorts generation, and a stress-free workflow for YouTube and TikTok.

PT-BR

How to Install and Run DeepSeek R1 for Free on Windows/Mac/Linux (Better than ChatGPT?)

Step-by-step guide to running DeepSeek R1 locally with Ollama and Open WebUI — no cloud, no monthly fees, full privacy. Includes a head-to-head performance comparison with ChatGPT.

PT-BR

How to Build an AI SDR: Qualifies and Books via Phone and WhatsApp

How to build a voice AI assistant that calls leads on WhatsApp and phone, qualifies them automatically, and books a closing meeting — built with VAPI, N8N, and OpenAI.

EN

Using CloudFormation to Deploy a Docker .NET 6 App to AWS ECS + Fargate

Automate the full AWS ECS infrastructure — ECR, Task Definition, Cluster, and Security Group — using a CloudFormation template instead of manual dashboard clicks.

EN

Deploy a Docker .NET 6 App to AWS using ECS

End-to-end walkthrough: create a .NET 6 project, Dockerize it, push the image to ECR, and run it on AWS ECS with a proper Task Definition and Security Group.

Career Timeline

Where I've built things

Brain and my current AI leadership work lead everything. Prior roles demonstrate range, resilience, and a consistent pattern of shipping at scale.

Jul 2025 — Present
8 Figure Agency
Santa Monica, CA · Remote

Head of AI & AI Architect

I lead the design and execution of Brain, our multi-tenant AI platform connecting to client tools (CRMs, ads platforms, Slack, meetings, data warehouses). My role sits at the intersection of product, architecture, and team leadership. I define the technical direction, set delivery standards, and manage a team of AI engineers and researchers. I set LLM strategy (model selection, prompting patterns, tool use, retrieval, guardrails), evaluation frameworks, and production observability with Grafana, Datadog, and LangSmith. The goal: grounding AI work in business impact, not demos.

PythonNext.jsLangChainCrewAIMCP ServerVector DBsMongoDBSupabasePostgreSQLAWSGrafanaDatadog
Jul 2022 — Jun 2025
KeHE Distributors
Naperville, IL · Remote

Tech Lead & Architect

Led the transformation of the e-commerce search platform for a NASDAQ-listed, $8B+ revenue food distributor. Guided the team through a POC, selected OpenSearch on AWS, and delivered a 14x search speed improvement and 2x sales conversion lift processing millions of daily queries. Also reduced B2B client onboarding from 15 minutes to 2 minutes. Managed multiple teams across retailer and back-office operations.

AWSOpenSearch.NET 8AngularDynamoDBSQSPostgreSQLDatadog
Nov 2013 — Aug 2024
Polen
Curitiba, BR & International

CTO & Co-Founder

Built the first social impact Open API in Latin America from scratch. Raised $550K in VC, scaled to 5,000+ companies and 25M+ users, donated $1.5M+ through the platform. Managed ~20 people, implemented OKR and Scrum across the entire org. Migrated from monolith to microservices (Docker, Kubernetes, Google Cloud). Served as CEO for 2 years before transitioning to CTO.

C# / .NETAngularMongoDBDockerKubernetesGoogle Cloud
Nov 2016 — Jul 2017
Bonuts
Curitiba, BR

CTO & Co-Founder

Built a photo mobile app (iOS + Android) that rewarded users for posting branded-frame photos — brands paid $0.10–$5.00 per approved post for social buzz and virality. Grew to 10K users before Instagram and Snapchat commoditized the space.

XamarinASP.NET / C#MongoDBAWSiOSAndroid
Jan 2012 — Aug 2012
Many To One Consulting
Curitiba, BR

Technical Lead & Senior Developer

Built a CMS serving 3 travel portals for B2W Group (Submarine, Shoptime, Americanas). ASP.NET MVC 4, SQL Server 2012, caching and geolocation. Led technical delivery as module lead on a Scrum team.

ASP.NET MVCSQL ServerC#Scrum
Jun 2011 — Dec 2011
BuscaPé
São Paulo, BR

Middle Developer

Anti-fraud system for one of Brazil's largest e-commerce platforms. Improved SQL database performance, introduced Cassandra for high-throughput scenarios, and raised the bar on unit and functional testing.

C#SQL ServerCassandraScrum
Jan 2010 — Apr 2011
FCamara Consulting & Training
São Paulo, BR

Middle Developer

Worked across Walmart.com.br (data layer optimization, MVC migration, jQuery, async), PRNewswire (first PM role — team coordination, Scrum, client contact), and Milevo (Facebook/Twitter/Google social integration).

ASP.NET / C#SQL ServerjQueryWeb Services
Production Systems

What I've shipped

Four production AI/ML systems, each from architecture to deployment. Real users, real data, real outcomes.

01

Brain

8 Figure Agency — Multi-Agent AI Platform

Multi-tenant, multi-agent OS connecting to CRMs, ads platforms, Slack, meetings, and data warehouses. Vector + graph + structured stores, LLM orchestration, reusable agent patterns, production observability. Whiteboard to multi-client production in months.

↓ 20-30 hrs / client / week
02

KeHE Search

OpenSearch B2B E-commerce · AWS

AI-powered search for a NASDAQ-listed food distributor handling millions of daily queries. Architected POC to production, created flexible architecture enabling Ads in search results.

14× speed · 2× conversion
03

Cognia

AI Psychologist MVP — Multi-Model Safety

Primary LLM for responses, fine-tuned models for personality and domain knowledge, smaller LLMs for toxicity screening, human-in-the-loop routing after 3 validation checks. End-to-end architecture to deployment.

Multi-model safety pipeline
04

TaskClaw

Open Source — AI Task Platform

Self-hostable Kanban + AI chat + knowledge management with bidirectional Notion/ClickUp sync. Public codebase demonstrating AI integration patterns.

github.com/DevOtts/taskclaw ↗
Technical Toolkit

What I work with

AI / ML / LLM

LangChainLangSmithCrewAIMCP ServerAnthropic ClaudeOpenAIMistralLLaMA / OllamaVector DBsKnowledge GraphsRAG PipelinesMulti-Agent ArchitecturesPrompt EngineeringFine-TuningN8N

Backend & Infrastructure

Python.NET / C#Node.jsPostgreSQLMongoDBSupabaseDynamoDBOpenSearchAWS (Lambda, SQS, S3)DockerKubernetesCI/CD

Frontend & Observability

Next.jsReactTypeScriptAngularGrafanaDatadogLangSmithCloudWatchSonarQube

Leadership & Process

AI Strategy & RoadmapArchitecture Decision RecordsTeam Management (20+)Scrum / KanbanOKRSprint Planning & DoDStakeholder CommunicationTechnical Mentoring
Background

Education & Details

B.Sc. Computer Science

Universidade Federal do Paraná (UFPR) · 2007—2011

Microsoft Student to Business (S2B)

C# Program · 2008

Startup Accelerators

Start-Up Chile · DotForge Impact (UK) · Start You Up (Brazil)

Yunus & Youth Fellow

Social Business · Alumni · Global Fellowship

Award

Startup Weekend Winner

Location

Curitiba, Paraná, Brazil · Remote, open to global

Languages

Portuguese (Native) · English (Full Professional) · Spanish (Working)

Notice Period

2 weeks

Technical Content

Youtube Devotts & Otimiza_ai — AI systems, agent architectures, automation

Let's talk about your AI challenges

I'm looking for the right AI leadership role. If you're building agentic systems and need someone who ships, let's connect.

LinkedIn →

Fernando's AI Clone

Ask anything about my experience, projects, or approach

Hey there! 👋

I'm Fernando's AI clone. Before we chat, what's your name?