2389 Research

Jun 2025 – Sep 2025

Designed and deployed cloud-based backend microservices in C++ supporting AI-powered conversational agents, integrating LLM APIs with hybrid decision pipelines for reliable, production-grade AI systems.

Category Backend / AI Engineering

Role Software Engineer

Duration 4 Months

Location Chicago, IL

Overview

At 2389 Research, I worked as a Software Engineer building cloud-based backend microservices in C++ that powered AI-driven conversational agents. These systems handled both structured and unstructured customer interaction data at scale.

A core focus of my work was integrating LLM APIs using structured output schemas and tool-calling patterns to enable deterministic execution workflows — ensuring that AI-powered features behaved predictably in production environments.

What I Did

Cloud Backend Microservices

Designed and deployed cloud-based backend microservices in C++ supporting AI-powered conversational agents handling structured and unstructured customer interaction data.

LLM API Integration

Integrated LLM APIs using structured output schemas and tool-calling patterns to enable deterministic execution workflows.

Hybrid Decision Pipelines

Designed hybrid rule-based + LLM decision pipelines to reduce hallucination and improve response reliability.

Fallback Logic

Implemented fallback logic between LLM reasoning and deterministic validation layers to ensure robust and reliable system behavior.

Technologies

C++ Cloud Services Microservices LLM APIs AI/ML Tool Calling Structured Output Swift C# Frontend Engineering Fullstack Backend Engineering

MORE PROJECTS EXPLORE

More Works More Works More Works More Works More Works More Works More Works More Works

Reality Hack at MIT (03)

Carbonbusters (01)

Get Off My Channel! (04)