AI Agent Engineer

About the Job

Project Role: AI Agent Engineer
Company: Farmioc
Location: Bengaluru, Karnataka, India
Work Model: On-site
Employment Type: Full-time
Experience: 3–5 Years

About Farmioc

Farmioc is an agricultural data intelligence and commodity trading ecosystem focused on helping businesses and stakeholders across the agriculture value chain make better data-driven decisions.

Its platform brings together agricultural data, market intelligence, commodity trading, predictive analytics, and AI-powered decision-support solutions across products such as DataStore, PriceHub, MarketHub, AgriTools, and AGVISOR.

About the Role

We are looking for an AI Agent Engineer to design, develop, and improve production-grade AI agents and agentic workflows for agricultural data, market intelligence, commodity analytics, and decision-support applications.

You will build AI systems that understand user intent, use tools and APIs, retrieve relevant context, perform multi-step tasks, and generate grounded responses. The role focuses on agent orchestration, tool integration, context engineering, evaluation, and production AI reliability.

You will build on data and retrieval capabilities provided by Farmioc's data engineering and platform systems rather than independently owning the underlying data infrastructure.

What You'll DoAI Agent Development

• Design and develop AI agents for agricultural data, market intelligence, analytics, and decision-support use cases.

• Build agents that understand user requests and complete multi-step tasks using appropriate tools and services.

• Develop reliable workflows for task planning, tool execution, validation, and response generation.

• Integrate LLMs into production applications using appropriate prompts, context, tools, and guardrails.

• Choose deterministic tools, APIs, or conventional software approaches when an LLM or agent does not provide a meaningful benefit.

Agent Workflow & Orchestration

• Design agent workflows involving task decomposition, tool selection, execution, validation, and response generation.

• Implement state management, conditional workflows, retries, fallbacks, and error handling.

• Build agentic workflows using frameworks such as LangGraph, LangChain, or similar technologies.

• Improve agent reliability across different agricultural and commodity-related use cases.

• Develop reusable orchestration patterns that can support multiple Farmioc applications.

Tools, APIs & Integrations

• Build and integrate tools for databases, APIs, search, analytics, ML models, and internal Farmioc services.

• Implement function and tool calling with appropriate input validation and error handling.

• Enable agents to interact with Farmioc data services and analytical systems.

• Develop reusable tools that can be used across multiple agent workflows.

• Define clear permissions and boundaries for tools that access data or perform actions.

Agricultural Data & Context

• Build agents that use Farmioc's agricultural data and retrieval capabilities to answer domain-specific questions.

• Work with datasets, metadata, metrics, historical information, and market information provided by Farmioc's data systems.

• Ensure agents retrieve relevant context before generating analytical responses or recommendations.

• Work with agricultural, environmental, market, supply, demand, yield, weather, and commodity information where relevant.

• Ensure generated responses remain grounded in available data and clearly identify the source or basis of important conclusions.

Market Intelligence & Decision Support

• Develop AI workflows for agricultural commodity and market intelligence.

• Enable agents to analyze commodity prices, market trends, supply-demand conditions, historical data, and related market factors.

• Build workflows that combine information from multiple tools and data sources to answer complex analytical questions.

• Support decision-support use cases across commodities, markets, regions, and time periods.

• Ensure outputs clearly distinguish data-driven results, model-based estimates, and generated explanations.

Agricultural Safety & Validation

• Build validation and guardrails for AI-generated agricultural recommendations and decision-support outputs.

• Ensure agents can identify insufficient, conflicting, outdated, or regionally inappropriate data before generating recommendations.

• Prevent agents from presenting unsupported conclusions or uncertain information as definitive advice.

• Support escalation or human review for higher-risk agricultural recommendations where appropriate.

• Work with relevant agricultural or domain experts to develop evaluation cases for important decision-support workflows.

• Ensure recommendations are appropriate for the geographic, crop, weather, and data context available to the system.

RAG & Context Engineering

• Use Farmioc's retrieval capabilities to provide relevant context to AI agents.

• Develop context strategies for agricultural documents, structured data, market information, and domain knowledge.

• Work with embeddings, vector search, metadata filtering, and hybrid retrieval where appropriate.

• Optimize context selection for relevance, accuracy, latency, and token usage.

• Investigate retrieval-related failures and work with data engineers to improve the underlying data or retrieval capabilities.

Agent Evaluation & Quality

• Develop evaluation datasets and test cases for agent behavior and domain-specific tasks.

• Evaluate planning, tool selection, context usage, data accuracy, workflow execution, and final responses.

• Identify failure modes such as incorrect tool selection, irrelevant context, unsupported conclusions, incorrect data interpretation, and unsafe recommendations.

• Measure agent performance across representative agricultural and market-intelligence use cases.

• Improve agent behavior through systematic evaluation, testing, and iteration.

Performance & Production AI

• Optimize model usage, token consumption, latency, tool calls, and inference costs.

• Reduce unnecessary LLM calls and improve workflow efficiency without compromising response quality.

• Evaluate model, prompt, context, and tool-selection strategies for different use cases.

• Deploy and maintain AI agents in production environments.

• Implement logging, monitoring, tracing, error handling, and operational safeguards.

• Troubleshoot production issues across agents, models, APIs, databases, retrieval systems, and external services.

Engineering & Product Collaboration

• Work closely with data, backend, platform, product, and other engineering teams.

• Translate product requirements and agricultural use cases into practical agent workflows.

• Contribute to architecture, technical design, testing, documentation, and production releases.

• Build reusable AI components that can be applied across Farmioc products.

• Clearly communicate agent limitations, evaluation results, and production risks to relevant stakeholders.

What We're Looking For

• 3–5 years of experience in AI, software engineering, machine learning, Generative AI, or a related field.

• Strong Python programming and software engineering fundamentals.

• Practical experience building LLM-powered applications and AI agents.

• Understanding of agentic workflows, tool calling, function calling, and workflow orchestration.

• Experience with prompt engineering and context engineering.

• Experience working with APIs, databases, and backend services.

• Understanding of RAG, embeddings, vector databases, and information retrieval.

• Experience evaluating and improving AI application reliability.

• Strong analytical and problem-solving skills.

• Ability to work effectively with cross-functional engineering and product teams.

Preferred Qualifications

• Experience with LangGraph, LangChain, LlamaIndex, or similar agent frameworks.

• Experience building multi-step or multi-agent systems.

• Experience with MCP or other AI tool-integration protocols.

• Experience with NL-to-SQL or AI-powered data analytics.

• Experience with agricultural data, commodity markets, market intelligence, or other domain-specific analytics.

• Familiarity with ML models, forecasting, predictive analytics, or recommendation systems.

• Experience with AI evaluation frameworks and observability.

• Experience optimizing LLM calls, token usage, latency, and inference costs.

• Experience with PostgreSQL, MySQL, MongoDB, or analytical databases.

• Experience with Docker, cloud and deployment platforms, CI/CD, and production deployments.

• Familiarity with data visualization and dashboard-based analytics.

Key Skills

Python | LLMs | Generative AI | AI Agents | Agentic Workflows | Prompt Engineering | Context Engineering | Tool Calling | Function Calling | RAG | Embeddings | Vector Databases | MCP | AI Evaluation | APIs | SQL | Agricultural Data | Commodity Analytics | Market Intelligence | Docker | Cloud

What We Offer

• Opportunity to build production AI systems for the agriculture and commodity intelligence domain.

• Work with real-world agricultural datasets and market intelligence use cases.

• Exposure to LLMs, AI agents, RAG, predictive analytics, and AI-powered decision-support systems.

• Opportunity to contribute to Farmioc's data, analytics, marketplace, and AI products.

• Opportunity to solve practical AI problems with direct business impact.

Why Join Farmioc

At Farmioc, you will work at the intersection of AI, agricultural data, market intelligence, and commodity trading. You will build AI systems that work with real-world data and support complex analytical and decision-making workflows across the agriculture value chain.

Equal Opportunity

Farmioc is committed to providing equal employment opportunities and maintaining an inclusive and professional workplace.