AI Software Engineer

Job highlights

• Experienced AI Software Engineer with strong Python backend skills, familiar with AI agents, Model Context Protocol, workflow orchestration, and open-source LLM integration
• Design, develop, test, deploy AI-enabled software solutions including agentic workflows, API integration, data pipelines, LLM routing, and workflow orchestration

About the Role

We are looking for a hands-on AI Software Engineer who can design, develop, test and deploy reliable AI-enabled software solutions. The engineer will work across Python backend development, AI agents, Model Context Protocol, workflow orchestration, data pipelines, databases, open-source LLMs, Docker and AWS infrastructure.

This is a software-engineering and delivery role. It is not limited to prompt writing, chatbot configuration, model experimentation or creating demonstrations. The selected candidate must be able to take a technical requirement from discussion through architecture, prototyping, implementation, testing, deployment, monitoring and documentation.

The role covers API integration, data parsing, tool execution and automated workflows. Detailed business requirements will be shared with shortlisted candidates.

Role Objective

The objective is to develop production-ready AI solutions that connect with external systems, process information, invoke tools, coordinate multiple steps and return measurable, traceable outputs.

Strong Python/backend ability is mandatory, while some specialised technologies can be learned on the job.

Key Responsibilities ( Some of them as per skills )

AI Agents and Agentic Workflows

• Design agents for research, data collection, parsing, classification, extraction, summarization, analysis and decision-support tasks.
• Build agents that call approved APIs, query databases, process files and invoke software tools through clearly defined interfaces.
• Add safeguards for hallucinations, unsupported conclusions, malformed tool inputs, incomplete information and tool failures.
• Build evaluator or critic stages for important outputs and provide human review for sensitive actions.

Data Collection, Parsing and Integration

• Integrate REST APIs, webhooks, authenticated services and third-party platforms.
• Build pipelines to ingest and normalize JSON, XML, CSV, Excel, HTML, PDF and plain-text information.
• Develop compliant web-data collection or browser-automation components when required.
• Handle pagination, rate limits, authentication expiry, duplicate records, missing fields and changing source formats.

Model Context Protocol

• Build and maintain MCP servers and clients.
• Expose approved APIs, databases, files and services as controlled tools with defined input and output schemas.

LLM Integration, Routing and Evaluation

• Work with open-source models from Hugging Face and other established repositories.
• Implement LLM routing based on task complexity, accuracy, privacy, latency, availability and cost.
• Configure fallbacks so workflows can continue when a model or provider is unavailable.
• Add rate-limit handling, timeouts, retries, caching and token-budget controls.

Workflow Orchestration

• Design stateful, resumable and observable workflows.

Work with LangGraph, Temporal, Celery, Airflow, Dagster or an equivalent technology.

• Persist workflow state so interrupted jobs can resume at the correct stage.

We do not expect applicants to be experts in every framework named in this posting. Strong Python and backend foundations, evidence of completing real software projects and the ability to learn quickly are more valuable than listing many tools without practical depth.

Open-Source Software Capability

The candidate should be comfortable finding, evaluating, installing and configuring open-source projects. They should be able to read documentation and source code, assess maintenance quality, resolve operating-system and package conflicts and judge whether a component is suitable for production.

How to Apply

Submit your updated resume, GitHub or portfolio link and details of one AI or backend system you personally developed. Briefly explain its architecture, deployment, your contribution and one difficult technical failure you resolved.

Include your current location, notice period, current compensation and expected compensation. Shortlisted candidates may complete a time-boxed practical exercise involving Python, an external API, an LLM tool call, database storage and Docker deployment.