Compensation: ₹8L – ₹12L • No equity

We are looking for someone who can design and build the intelligence layer powering our next-generation education platform.

You will work across LLMs, RAG, AI agents, model routing, personalization, evaluation, AI infrastructure, and ML/data intelligence.

This is a hands-on engineering role for someone who wants to build production-grade AI systems rather than simply integrate APIs.

You will own important architectural decisions and help create AI systems that are scalable, cost-efficient, reliable, personalized, and flexible across multiple foundation-model providers.

Requirements

• Strong understanding of LLM application architecture and modern Generative AI systems
• Hands-on experience with OpenAI, Anthropic, Google Gemini, or other foundation-model APIs
• Experience building production applications using LLMs, structured outputs, tool/function calling, and prompt/system design
• Strong understanding of RAG architectures, embeddings, vector databases, semantic search, hybrid search, chunking, reranking, and retrieval pipelines
• Experience designing AI orchestration, model routing, context management, and multi-step AI workflows
• Strong Python programming and software engineering fundamentals
• Understanding of APIs, asynchronous processing, databases, caching, queues, and scalable backend systems
• Ability to evaluate AI systems based on quality, accuracy, latency, reliability, token usage, scalability, and cost
• Understanding of AI evaluation, hallucination detection, grounding, output validation, and reliability mechanisms
• Ability to design systems that remain flexible as models, providers, requirements, and product usage evolve
• Strong understanding of when to use LLMs and when deterministic code, databases, search, rules, or traditional ML would be more appropriate
• Ability to make engineering trade-offs and choose the simplest reliable architecture for a given problem

Responsibilities

• Design and own Novevex's AI architecture across LLM integration, AI orchestration, model routing, context management, RAG, vector search, knowledge systems, AI agents, personalization, evaluation, safety, reliability, and observability
• Build production-grade AI applications on top of OpenAI, Anthropic, Google Gemini, and other suitable foundation models
• Design vendor-flexible model architectures that allow Novevex to switch between models and providers without redesigning the entire platform
• Evaluate and select models based on quality, capability, latency, reliability, token usage, infrastructure requirements, and cost
• Design cost-effective AI architectures using techniques such as model routing, caching, retrieval, batching, precomputation, and appropriate model selection
• Build scalable AI systems capable of supporting growth from hundreds of students to hundreds of thousands or millions of users
• Design and implement RAG and knowledge-intelligence systems including document ingestion, chunking, embeddings, vector search, hybrid retrieval, metadata filtering, reranking, grounding, and context pipelines
• Build personalized AI experiences using student profiles, class, board, subject, chapter, topic, performance, learning history, previous interactions, weaknesses, and current learning objectives
• Develop intelligent AI workflows that combine student context, academic knowledge, retrieval, model selection, generation, validation, and personalization
• Build AI evaluation pipelines to measure correctness, relevance, hallucination, grounding, consistency, explanation quality, question quality, difficulty, and structured-output validity
• Implement reliability mechanisms including validation, guardrails, fallback models, error handling, monitoring, and quality checks
• Design AI observability systems to monitor model performance, latency, token consumption, costs, failures, and user-level outcomes
• Work with student-performance data and help design the interface between LLM-based intelligence and future ML/data-science systems
• Collaborate with backend, product, and data teams to integrate AI capabilities into the broader Novevex platform
• Continuously research and evaluate new AI models, frameworks, architectures, and techniques that can create meaningful product advantages
• Make architecture decisions based on business requirements, expected scale, reliability, maintainability, performance, and cost rather than technology trends alone

Technical Skills

Required

LLM / Generative AI

• OpenAI / Anthropic / Gemini APIs
• LLM application architecture
• Context engineering
• Structured outputs
• Tool/function calling
• Prompt/system design
• Model evaluation

RAG / Retrieval

• Embeddings
• Vector databases
• Semantic search
• Hybrid search
• Metadata filtering
• Reranking
• Document pipelines

AI Engineering

• Python
• API development
• Asynchronous workflows
• Caching
• Retries/fallbacks
• Observability
• Performance optimization

Architecture

• Scalable system design
• API/service architecture
• Cloud infrastructure concepts
• Distributed systems fundamentals
• Security and reliability
• Cost optimization

Strongly Preferred

• AI agents / agentic workflows
• Multimodal AI
• Recommendation systems
• Personalization engines
• Knowledge graphs / concept graphs
• ML fundamentals
• Model routing
• Inference optimization
• AI experimentation
• Evaluation frameworks
• Production AI systems

What We Are Not Looking For

• A prompt engineer who only writes prompts
• Someone who only connects an LLM API
• An academic ML researcher who only wants to train models
• Someone focused on building a foundation LLM
• Someone who chooses technology based only on popularity
• Someone who ignores API/infrastructure cost
• Someone who builds a demo but cannot design a production system