phoenix
phoenixPhoenix - Your AI Assistant
Inactive session will be ended in 24 hours
End Session

AI Agent Developer - LLM Tools, Support Bots & Agent Workflows

  • Posted:
  • Proposals: 15
  • Remote
  • #4502023
  • Open for Proposals
Aurel H.Muhammad A.Pavlo H.CYTNHIA W.Chintan P. + 10 others have already sent a proposal.
  • 0

Description

Experience Level: Entry
Looking for an experienced Python developer with hands-on AI/agent work (LLM tool-calling, RAG, LangChain/LangGraph or similar) to support ongoing development, fixes, and small features. Remote, async-friendly.

About the work
I need a reliable developer to support me on AI-related Python projects. Work is mostly backend and agent logic — not generic web CRUD.

Typical tasks:
- Build and extend AI agents that call tools (APIs, mocked services, email, refunds, etc.)
- Implement agent loops (plan -> tool call -> observe -> next step)
- Integrate LLMs (OpenAI or similar) with structured tool schemas, or rule-based flows where appropriate
- Work with mocked external systems during development (no production secrets in dev)
- Write clear README, basic tests, and honest notes on technical debt
- Fix bugs, refactor small modules, and help prepare code for review or demos
- Optional: FastAPI*endpoints to expose agents as APIs

You will work from existing repos or small greenfield scripts. I value readable code, clear structure, and someone who can explain trade-offs.


Required experience
Please only apply if you can point to **real examples** (GitHub, portfolio, or PPH history) of most of the following:
| **Python** | 3+ years; clean modules, typing, error handling |
| **AI agents** | Tool-calling agents, not only chatbots |
| **LLM integrations** | OpenAI API or equivalent; prompts + tool definitions |
| **Frameworks** | LangChain, LangGraph, CrewAI, or solid **vanilla Python** agent loops |
| **RAG** (nice) | Retrieval, chunking, vector search for Q&A bots |
| **APIs** | REST/FastAPI; calling third-party APIs from agent tools |
| **Communication** | Good English; short written updates; ask when requirements are unclear |

Nice to have
- Customer support / workflow automation agents (tickets, email, refunds, CRM)
- **Authorization and risk thinking** for agents (who can trigger money actions, escalation paths)
- **Runtime governance** or guardrails (policy checks before high-impact steps)
- Docker, basic CI, pytest
- Experience in **startup / solo-engineer** environments (ownership, not only tickets)

New Proposal

Login to your account and send a proposal now to get this project.

Log in

Clarification Board Ask a Question

  • CYTNHIA W.

    • Which agent framework are you primarily using today—LangGraph, LangChain, CrewAI, OpenAI Assistants, or mostly custom Python implementations?

    • Are the current projects primarily customer-support agents, internal workflow agents, research agents, or a mix of several use cases?

    • How much of the workload involves extending existing repositories versus building new agent workflows from scratch?

    • Do your current agents use RAG/vector search, or is most context provided directly through tools and APIs?

    • What is the typical engagement volume per week in terms of hours and ongoing workload?

    • Are you looking for someone to participate in architecture decisions, or mainly implementation and maintenance work?

    • Do you have an existing testing strategy for agent behavior, tool execution, and guardrails, or would that be part of the role?

    • Which LLM providers are currently in use (OpenAI, Anthropic, Gemini, local models, etc.)?

    • Are deployments handled locally, through Docker, or on cloud platforms such as AWS, GCP, Azure, Railway, or Render?

    • What would a successful first month look like for the developer you hire?