Company: CodePraxis

Experience: 2-3 Years

Role: AI Engineer, Agentic Systems

Employment Type: Full time

About the Role

At CodePraxis, we're building products where AI agents do real work. As an AI Engineer, you'll own the design and delivery of our AI agents: systems that plan, reason, call tools, and complete multi-step tasks on their own. You'll work directly with the founders and engineering team, with real ownership of what ships.

Key Responsibilities

- Design, build, and deploy production AI agents that plan, use tools, and complete multi-step workflows

- Build multi-agent systems with clear orchestration, hand-offs, and shared state

- Develop tool and function-calling integrations that connect agents to our APIs, databases, and third-party services

- Implement memory and context management (short-term, long-term, RAG) so agents stay accurate across long tasks

- Build RAG pipelines using embeddings, vector databases, and semantic search

- Create evaluation frameworks for agents: task success rates, regression tests, trajectory analysis, and guardrails

- Optimize prompts, latency, and cost across LLMs such as Claude, GPT, and Llama

- Add observability, tracing,



and error handling so agent behavior is debuggable and reliable in production

- Build safety controls: output validation, permission boundaries, and human-in-the-loop checkpoints

- Collaborate with product and engineering to turn real user problems into agent-driven features

Requirements

- 2-3 years of experience in software engineering or AI/ML, with hands-on experience building LLM-based applications

- Strong Python skills and solid API/backend fundamentals (REST, async, queues)

- Practical experience building AI agents: tool use, planning, multi-step workflows, and orchestration

- Experience with frameworks such as LangChain, LangGraph, LlamaIndex, or the Claude Agent SDK

- Experience with RAG, vector databases, and embeddings

- Familiarity with cloud platforms (AWS, Azure) and deploying AI systems to production

- A rigorous approach to evaluating and debugging non-deterministic systems

- Strong problem-solving skills and the ability to work independently

Nice to Have

- Experience with MCP, agent observability tools (LangSmith, Langfuse), or fine-tuning

How to Apply

- If you've built AI agents that actually work in production, apply with your resume.

📌 AI Engineer (Bhubaneswar)
🏢 codepraxis
📍 Bhubaneswar