Dieture AI Developer Hiring 2026: Remote Job with ₹10–12 LPA Salary
Dieture is hiring an AI Developer to build customer-facing artificial intelligence applications that can understand conversations, retrieve relevant information, interact with software tools and deliver reliable responses. The vacancy lists compensation of ₹10 lakh to ₹12 lakh per year and allows remote hiring from locations worldwide, according to the job listing.
The role is focused on production-ready AI applications rather than research alone. The selected developer will work with Python, large language models (LLMs), retrieval-augmented generation (RAG), APIs, conversational memory, evaluation systems and safeguards that help AI applications behave reliably in real-world use.
This opportunity is particularly relevant to software engineers and AI developers who have already built backend services or integrated LLM APIs and want to work on more advanced conversational AI systems.
What will you actually build?
The role description centres on a practical challenge: turning a user’s request into an accurate, contextual response while allowing the AI application to retrieve information and interact with other systems safely.
For example, a conversational AI application might need to understand a customer’s question, retrieve relevant information from a knowledge base, call an approved API, maintain context from an earlier message and return a structured response. Dieture wants an engineer who can implement and maintain this complete workflow.
The main areas of work include:
- LLM application development: Build conversational and generative AI features using modern language models.
- Python backend services: Develop and maintain APIs using FastAPI or a similar framework.
- RAG pipelines: Implement embeddings, vector search, metadata filtering, reranking and contextual retrieval.
- Memory and personalisation: Design short-term and long-term mechanisms for retaining relevant conversational context.
- Tool integrations: Connect AI applications to internal APIs, databases and third-party services.
- AI workflows: Develop agentic capabilities where they are useful for the application.
- Production reliability: Implement retries, timeouts, model fallbacks and appropriate error handling.
- Monitoring and evaluation: Track response quality, latency, errors, token usage and cost, then improve behaviour based on production observations.
The job description does not identify a particular AI product or customer use case that the successful candidate will own first. Candidates should therefore prepare to discuss their engineering approach rather than assume the work will involve one specific chatbot or framework.
Technical skills that matter most
Python, FastAPI and API development
Strong Python programming is a core requirement. Experience with FastAPI or a comparable backend framework is also expected, along with REST API integration, asynchronous programming, Git and standard software development practices.
Candidates should be comfortable writing maintainable, testable code and debugging problems that span application logic, APIs and external services.
RAG, embeddings and retrieval
Dieture expects an understanding of retrieval-augmented generation and embeddings. RAG systems retrieve relevant information to provide an LLM with context before it generates a response.
Preparation should cover how documents are prepared and indexed, how vector search works, why metadata filtering can matter, and how reranking may improve retrieval relevance. It is also useful to understand that poor retrieval can lead to poor answers even when the underlying model is capable.
Prompt engineering, structured outputs and tool calling
The position goes beyond writing prompts. It includes context engineering, structured generation, streaming and function or tool calling.
Candidates should be able to explain how they make model outputs usable by software, validate returned data, handle failed tool calls and prevent a model from invoking actions it is not authorised to perform.
AI evaluation and production debugging
An AI application can return different answers to similar requests, so conventional unit tests alone may not be sufficient. Dieture specifically mentions automated evaluation and regression testing for AI behaviour.
The role also requires monitoring response quality, latency, errors, token consumption and cost. Engineers should understand how to identify regressions, investigate unexpected outputs and balance response quality against performance and operating expenses.
AI safety is part of the engineering work
Dieture explicitly includes safety and reliability in the responsibilities. The developer will contribute to safeguards such as:
- Input and output validation.
- Protection against prompt injection.
- Limits on which tools an AI system can access or invoke.
- Personally identifiable information (PII)-aware data handling.
- Domain-specific guardrails and safe fallback responses.
- Human escalation when automated handling is unsuitable.
- Monitoring and auditability.
These requirements make the position different from a role focused only on connecting an LLM API to a basic interface. The goal is to build a system that remains useful and controlled when users provide unexpected inputs, retrieval fails or the model produces an unreliable answer.
Which additional tools could help?
The job description lists several technologies as desirable rather than mandatory.
| Area | Examples mentioned |
|---|---|
| Model providers | OpenAI, Anthropic, Google Gemini |
| AI orchestration | LangChain, LangGraph |
| Model routing | OpenRouter |
| Memory | Mem0 |
| AI observability | Langfuse, LangSmith |
| Databases and caching | PostgreSQL, MongoDB, Redis |
| Vector search | Vector databases |
| Infrastructure | Docker, Kubernetes, AWS, Azure, GCP |
| Asynchronous processing | Background workers and queues |
| Real-time applications | WebSockets and streaming responses |
You do not need to claim experience with every item. The description identifies these as useful additional skills, while the required skills focus on Python, backend development, LLM integration, RAG fundamentals and production engineering.
How to prepare a portfolio project
A strong demonstration project for this role would be a conversational assistant that retrieves information from a controlled knowledge base and can call an approved application API.
To make the project relevant, show the complete engineering process:
- Build a Python API that accepts user requests.
- Retrieve relevant information using embeddings and vector search.
- Generate structured model responses and validate their format.
- Add a tool call with explicit permissions and error handling.
- Test difficult cases, including missing information and prompt-injection attempts.
- Measure response quality, latency and token cost.
- Add logging and regression tests so future changes can be evaluated.
A project that documents its limitations and failure cases can demonstrate more practical engineering judgement than a simple chatbot demo. This is preparation advice based on the role requirements, not an additional condition imposed by Dieture.
Who should consider applying?
This position is best suited to candidates who have approximately 2–4 years of software or AI engineering experience and can demonstrate practical work with LLM applications.
It may be a good fit if you have built backend services, integrated model APIs, implemented retrieval pipelines or debugged AI applications beyond the initial prototype stage. Experience with customer-facing conversational AI is particularly valuable according to the listing.
Candidates should be comfortable combining several engineering disciplines: backend development, retrieval, application data, model behaviour, testing, security and production monitoring. The description does not present the role as primarily an ML research position.