Artificial Intelligence Engineer

We at Coforge are hiring for Agentic AI Engineer role with the following skillset

Design, develop, test, and deploy AI-agent workflows that use language models, enterprise data, APIs, and tools to complete business tasks safely and efficiently.

• Convert business requirements into agent workflows, decision logic, tool calls, and human approval steps.
• Build single-agent and basic multi-agent solutions.
• Develop prompts, structured outputs, routing, retries, memory, and exception handling.
• Implement RAG, grounding, vector search, and context management.
• Integrate agents with APIs, databases, enterprise applications, and workflow platforms.
• Test accuracy, task completion, safety, latency, cost, and hallucination risk.
• Implement guardrails for privacy, permissions, prompt injection, and responsible AI.
• Diagnose issues using logs, traces, model responses, and tool-call history.
• Develop maintainable Python or Type Script code and support CI/CD deployment.
• Create technical documentation, workflow diagrams, and operational runbooks.

• Python and/or Type Script
• Generative AI, large language models, prompt engineering, and structured outputs
• Agent frameworks such as Lang Graph, Semantic Kernel, Auto Gen, or CrewAI
• RAG, embeddings, vector databases, and enterprise search
• REST APIs, SQL, Git, Docker, cloud services, and CI/CD
• AI evaluation, monitoring, security, and responsible-AI practices

• Bachelor’s degree in Computer Science, Engineering, IT, Data Science, or equivalent experience.
• 2–5 years in software, data, automation, or AI engineering.
• Experience building at least one generative-AI, RAG, chatbot, automation, or agentic-AI solution.
• Experience integrating applications through APIs.
• Strong problem-solving and analytical skills
• Experimentation and continuous-learning mindset
• Good engineering and documentation discipline
• Ability to collaborate with technical and business teams
• Ownership of development, deployment, and issue resolution
• Sound judgement when handling AI risks and limitations

• Agent accuracy and task-completion rate
• Business adoption and productivity improvement
• Reliability, latency, and cost efficiency
• Defect and hallucination rates
• Delivery against sprint and release commitments
• Code quality, test coverage, and documentation