Job description
Job Purpose We are seeking a dynamic AI Engineer to join our pioneering Agentic AI team.
The ideal candidate will possess a strong foundation in delivering projects on both SaaS & PaaS based platforms, Prompt Engineering, UI development and development of APIs.
You will be involved in end-to-end stage of development life cycle.
Duties and Responsibilities Deliveries with respect to Agentic AI Platform: · Build and maintain autonomous software agents using state-of-the-art LLM frameworks.
· Collaborate with product owners and domain experts to build reusable components for business process automation.
· Develop core infrastructure and reusable components to support the deployment of agent-based AI systems.
· Work on agent orchestration, prompt engineering, and LLM-powered integrations.
· Implement scalable solutions integrated with CRM systems and enterprise data platforms.
· Contribute to the design of modular, extensible, enterprise-grade architecture.
· Fine-tune and evaluate AI agents for speed, accuracy, performance and maintainability across business units.
· Contribute to CI/CD automation and maintain operational stability of agent services.
Generative AI & Model Optimization: · Writing and Optimizing Prompts · Fine-tune LLMs/SLMs with proprietary NBFC data.
· Perform distillation, quantization of LLMs for edge deployment.
· Evaluate and run LLM/SLM models on local/edge server machines.
Self-Learning Frameworks: · Build self-learning systems that adapt without full retraining (e.
g., learn new rejection patterns from calls).
· Implement lightweight local models to enable real-time learning on the edge.
Key Decisions / Dimensions Platform Design & Delivery, Model Selection, Customization (if any) & Testing: · Choosing the right model for various agentic and autonomous actions.
· Selecting appropriate model so that Agents can complete the autonomous tasks in efficient manner.
· Defining reusable components in the platform.
Delivery using configuration approach should be first preference, in case that does not work should go for customization.
· Defining configuration parameters and incorporate them as platform design.
· Testing and end to end testing of the project deliverable.
· Load balancing between different models.
· Always have switch on/switch off feature.
· Must have all services backed up on primary/HA & DR servers.
Prompt Engineering: · Prompt Design & Development: Crafting prompts that guide AI systems to produce desired outputs for various applications, such as text generation, translation, question answering, and creative writing.
· Testing and Evaluation: Analysing the effectiveness of prompts and refining them based on results to ensure accurate and relevant responses.
· Bias Mitigation: Designing prompts that minimize bias and ensure fair and equitable outcomes from AI systems Major Challenges Support from other platform owners · Agents must learn from failed interactions · Building a Agents that doesn't just answer but negotiates with human-like reasoning.
· Running large AI models in low-latency, low-bandwidth environments without cloud dependency.
· Getting the end-to-end domain knowledge Managing data and information security of the agentic application.
Required Qualifications and Experience a) Qualifications · Bachelor’s or Master’s degree in Computer Science, Engineering, or a related field.
· Experience: 6–8 years of experience in AI/ML, with exposure to Python/Java, Node.
JS, JavaScript.
b) Work Experience · Strong understanding of prompt engineering, tool calling, and agent orchestration.
· Good understanding of Microsoft Foundry + MS CoPilot Studio/ Agentforce or Equivalent.
· Strong programming skills in languages such as Python, Node.
JS, JavaScript, React, Java, Postgres.
· Familiarity with LangChain, Semantic Kernel, CrewAI, or LangGraph.
· Understanding of Microsoft Copilot Studio, Power Automate, and Power BI · Familiarity with the Agent Development Kit (ADK) and protocols like MCP & A2A
This job post has been translated by AI and may contain minor differences or errors.