Floor 01 // Enterprise AI

Nexus Corporation

Senior AI Engineer
2024 - Present
San Francisco, CA
12 Eng Team

Nexus Corporation builds autonomous AI agents for Fortune 500 supply chains. I was hired to rip out their slow, hallucination-prone LangChain wrappers and build a deterministic, sub-second latency agentic orchestration engine from scratch.

Primary Stack
PythonFastAPIQdrantRedisDockerAWS ECSOpenAI
System Ownership

RAG Pipeline

RBAC Auth Service

Standard Execution Flow
Business Sync
Architecture
Implementation
Testing
Deployment
Optimization
Critical Engineering Problems Solved
A Typical Day
09:00 AM

Async Standup

Reading Slack threads. I avoid synchronous morning meetings to preserve deep work windows.

10:00 AM

Architecture Drafting

Writing RFCs (Request for Comments) for the new embedding model migration. No code is written without a peer-reviewed plan.

01:00 PM

Execution

Translating the RFC into Python. Building the Qdrant ingestion worker.

04:00 PM

Load Testing

Using Locust to hammer the new worker with 10k concurrent requests to find breaking points.

Engineering Pipeline
Figma
Design & Req
Git / GitHub
Version Control
FastAPI / Node
Backend Logic
PostgreSQL
Persistence
Docker / K8s
Containerization
AWS / Vercel
Deployment
What I learned here

LangChain is great for prototypes, but custom orchestration is required for production reliability.

Never trust an LLM to output valid JSON without strict Pydantic validation and retry logic in the loop.

Floor 02 // FinTech

DataFlow Inc

Backend Engineer
2022 - 2024
Remote
8 Eng Team

DataFlow processed millions of micro-transactions daily. When I joined, their monolithic Node.js backend was falling over during peak trading hours. I led the migration to a microservices architecture.

Primary Stack
Node.jsTypeScriptPostgreSQLKafkaKubernetesDatadog
System Ownership

Event Streaming

Database Sharding

Standard Execution Flow
Business Sync
Architecture
Implementation
Testing
Deployment
Optimization
Critical Engineering Problems Solved
A Typical Day
08:30 AM

Datadog Review

Checking error rates and latency spikes from the European market session.

10:00 AM

Pair Programming

Working with junior engineers to untangle legacy monolithic controllers.

01:00 PM

Migration Coding

Writing the Kafka producers for the new event-driven architecture.

05:00 PM

Canary Deployment

Routing 5% of production traffic to the new microservice and monitoring logs.

Engineering Pipeline
Figma
Design & Req
Git / GitHub
Version Control
FastAPI / Node
Backend Logic
PostgreSQL
Persistence
Docker / K8s
Containerization
AWS / Vercel
Deployment
What I learned here

Microservices don't solve bad code, they just distribute it. Fix the domain logic before splitting the monolith.

Always have a one-click rollback plan for database migrations.

Floor 03 // Logistics

TechNova

Software Engineering Intern
2021
Austin, TX
5 Eng Team

My first taste of production engineering. I was tasked with building an internal dashboard to track fleet vehicles in real-time, learning how to handle raw WebSockets and messy geospatial data.

Primary Stack
ReactExpress.jsMongoDBSocket.ioGoogle Maps API
System Ownership

Real-Time Telemetry

Route Optimization

Standard Execution Flow
Business Sync
Architecture
Implementation
Testing
Deployment
Optimization
Critical Engineering Problems Solved
A Typical Day
09:00 AM

Mentor Sync

Reviewing yesterday's code and asking questions about JavaScript closures.

10:30 AM

Frontend Hacking

Struggling with React state management to make the map markers move smoothly.

02:00 PM

Backend Setup

Writing Express routes and learning how to use Postman to test them.

04:30 PM

Code Review

Getting my PR shredded by the senior dev, learning more in 30 minutes than in 3 years of college.

Engineering Pipeline
Figma
Design & Req
Git / GitHub
Version Control
FastAPI / Node
Backend Logic
PostgreSQL
Persistence
Docker / K8s
Containerization
AWS / Vercel
Deployment
What I learned here

Indexes matter. A missing index will take down your entire application.

Don't be afraid to ask for help when you are stuck for more than 2 hours.