Posted 1mo ago

Agent Dev Engineer (Full-Time/Intern)

@ Flowith
San Francisco, California, United States
OnsiteFull Time, Internship
Responsibilities:Design reasoning, Build orchestration, Evaluate performance
Requirements Summary:Proficient in Python; strong experience with LLM ecosystems, frameworks like LangChain/AutoGen; expert in prompt engineering; full-stack capability; problem solving; experience with RAG/vector databases; internship or GitHub AI projects valued.
Technical Tools Mentioned:Python, LangChain, LangGraph, AutoGen, FastAPI, Node.js, OpenAI API, Claude, Vector Databases, APIs
Save
Mark Applied
Hide Job
Report & Hide
Job Description

Role Overview

You are the "Architect of Intelligence" at Flowith. In this role, you design the reasoning brains behind our product. You will bridge the gap between raw Large Language Models and complex, real-world workflows by building sophisticated AI Agents. Your mission is to evolve Flowith Agent, transforming static algorithms into dynamic partners that can reason, plan, and execute tasks alongside our users.

Key Responsibilities

  • Agent Core Development: Design and implement the core reasoning logic for Flowith Agents, enabling them to handle complex planning, tool invocation, and long-term memory management.
  • Workflow Orchestration: Build and optimize multi-agent collaboration frameworks that allow AI to navigate the Flowith Canvas, interacting with various nodes and external APIs seamlessly.
  • Evaluation & Optimization: Establish rigorous evaluation benchmarks for Agent performance. Systematically analyze "hallucinations" or reasoning failures and iterate on prompts and logic to ensure reliability.
  • Tooling & Integration: Develop and maintain "skills" (tools) for Agents, expanding their capabilities to interact with the web, databases, and third-party SaaS platforms.
  • Front-Frontier Exploration: Stay at the cutting edge of AI research (e.g., new model capabilities, agentic patterns) and rapidly prototype features that keep Flowith at the forefront of the AI-native era.