About Opportunity

Axsot is hiring a Generative AI Engineer – Agentic AI for its portfolio company, Faxoc.

Faxoc develops evidence-based executive, governance, investment and enterprise decision intelligence for Boards, CXOs, Private Equity firms and institutional investors.

Faxoc’s products combine structured client intake, controlled evidence, analytical methodology, risk ratings and institutional reporting. The successful candidate will design and build AI systems that can reason across structured and unstructured information, retrieve relevant evidence, execute defined workflows and generate reliable intelligence within controlled product and methodology boundaries.

This is not a conventional chatbot or prompt-engineering role. It is an engineering position at the intersection of Generative AI, agentic systems, LLMs, data, research intelligence and enterprise decision-support applications.

Role Purpose The Generative AI Engineer will design, develop and deploy AI agents and LLM-powered systems that support Faxoc’s intelligence products.

The role will focus on ensuring that AI systems can:

- Retrieve and reason over relevant evidence and structured data.
- Execute controlled multi-step analytical workflows.
- Use approved tools, APIs and data sources appropriately.
- Maintain context, state and traceability across agent workflows.
- Generate structured, evidence-grounded intelligence.
- Distinguish facts, evidence, analysis, assumptions and conclusions.
- Operate within defined methodology, security and governance controls.
- Produce consistent and reliable outputs suitable for institutional decision-making.

Agentic AI Development

- Design and build AI agents capable of executing multi-step research and analytical workflows.
- Develop agentic workflows using frameworks such as LangGraph, LangChain, CrewAI, AutoGen or equivalent technologies.
- Implement planning, reasoning, tool selection, execution and verification workflows.
- Design single-agent and multi-agent architectures where appropriate.
- Build controlled agent workflows rather than unconstrained autonomous behaviour.
- Implement state management, memory and context handling across complex workflows.
- Define appropriate human-in-the-loop intervention and approval points.
- Ensure agents fail safely when required information or evidence is unavailable.

LLM and Generative AI Engineering

- Develop production applications using commercial and/or open-source Large Language Models.
- Design structured prompting, system instructions and output schemas.
- Implement function calling and tool-use capabilities.
- Build LLM pipelines for classification, extraction, summarization, analysis and intelligence generation.
- Evaluate different models for accuracy, latency,



reliability and cost.
- Implement techniques to reduce hallucination, unsupported inference and inconsistent outputs.
- Develop structured outputs suitable for downstream analytical and reporting systems.

Evidence, RAG and Knowledge Systems

- Design and implement Retrieval-Augmented Generation (RAG) systems.
- Connect LLMs with controlled internal knowledge bases, evidence repositories and structured datasets.
- Work with vector databases, embeddings, metadata and hybrid retrieval approaches.
- Ensure retrieved information is relevant, attributable and appropriately scoped.
- Preserve source references and evidence traceability throughout AI workflows.
- Handle conflicting, incomplete, outdated or unavailable information appropriately.
- Prevent AI systems from presenting unsupported information as established fact.

Research and Intelligence Workflows

- Build AI workflows that support executive, governance, investment and enterprise research.
- Automate structured research, information extraction and evidence synthesis.
- Develop agents capable of gathering, organizing and analyzing information from approved sources.
- Convert complex information into structured analytical outputs.
- Support risk identification, thematic analysis, comparison and intelligence generation.
- Build verification stages to distinguish retrieved evidence from AI-generated interpretation.
- Ensure analytical outputs remain aligned with approved Faxoc methodologies.

AI Quality, Evaluation and Reliability

- Establish evaluation frameworks for LLM and agentic AI performance.
- Create test datasets, expected outputs and evaluation criteria for AI workflows.
- Measure accuracy, relevance, consistency, hallucination, latency and failure rates.
- Test edge cases, contradictory information, missing evidence and ambiguous inputs.
- Implement automated and human evaluation mechanisms.
- Monitor AI behaviour following model, prompt, data or workflow changes.
- Develop regression testing for critical AI workflows.
- Continuously improve agent reliability and output quality.

AI Architecture and Integration

- Integrate AI agents with Faxoc’s existing applications, APIs, databases and data pipelines.
- Build production-ready services using Python and appropriate backend technologies.
- Develop APIs and services for agent execution and LLM-powered functionality.
- Implement authentication,



authorization and controlled access to AI tools and data.
- Work with cloud infrastructure, containers and deployment pipelines.
- Optimize AI applications for scalability, performance and cost.
- Maintain clean, modular and maintainable production code.

Security and Governance

- Ensure AI agents access only authorized data, tools and systems.
- Implement controls against prompt injection, data leakage and unauthorized tool execution.
- Prevent confidential or restricted information from being unintentionally exposed through AI outputs.
- Maintain appropriate logging and auditability of significant AI actions.
- Implement safeguards for sensitive enterprise and institutional information.
- Ensure AI-generated conclusions remain distinguishable from verified evidence.
- Escalate methodology, security or data-integrity issues where AI behaviour could materially affect decision outputs.

Product Collaboration

- Work closely with product, research, data and engineering teams to identify opportunities for AI automation.
- Translate business and intelligence requirements into practical AI architectures.
- Identify processes where agentic AI can improve research quality, speed and decision usefulness.
- Prototype new AI capabilities and move validated solutions into production.
- Balance innovation with reliability, explainability, security and commercial value.
- Contribute to the evolution of Faxoc’s AI and intelligence platform.

Engineering and Technical Requirements

- Strong proficiency in Python.
- Hands-on experience developing Generative AI / LLM applications.
- Practical experience building AI agents or agentic workflows.
- Experience with LLM APIs, function/tool calling and structured outputs.
- Strong understanding of RAG, embeddings, vector databases and retrieval systems.
- Experience with LangGraph, LangChain, CrewAI, AutoGen or comparable frameworks.
- Understanding of APIs, databases, backend services and cloud deployment.
- Experience with Git, Docker and production software development.
- Strong understanding of AI evaluation, reliability and hallucination mitigation.
- Ability to build, test, deploy and maintain production-grade AI systems.

Preferred Experience

- Experience building multi-agent systems.
- Experience with MCP or similar tool/context integration protocols.
- Experience with AI observability and evaluation platforms.
- Experience working with enterprise, financial, investment, governance or research applications.
- Experience integrating structured and unstructured data into AI workflows.
- Experience taking AI prototypes from concept to production.
- Solid understanding of information security and responsible AI practices.

📌 Generative AI Engineer (India)
🏢 AXSOT
📍 India