AI Research Engineer: Reasoning & Retrieval

Home/Jobs/AI Research Engineer: Reasoning &

• Retrieval

AI Research Engineer: Reasoning &

• Retrieval

Auric AI

Mumbai / Bengalore

3-5 years

Today

$24.1K–37.3K/yr

Full-time

Onsite

Skills Required

RAG

Information Retrieval

Machine Learning

NLP

Description

Auric AI is building a reasoning system over millions of messy, multilingual intelligence documents, fully hosted in India on infrastructure they control. They are hiring one research engineer to own the architecture for reasoning and retrieval.

Company: Auric AI

Role: AI Research Engineer: Reasoning &

• Retrieval

Location: Bengaluru/Mumbai (Onsite)

Experience

• No experience requirement
• Built something that survived real, messy data
• Can reason about language models and retrieval mechanically
• Can explain why naive RAG fails on multi-hop temporal questions

Qualification

• No degree requirement

Responsibilities

• Own the architecture for a retrieval and reasoning system
• Design retrieval that can bound its own recall
• Build reasoning across many hops and sources
• Surface contradictions rather than averaging them away
• Propagate confidence explicitly through multi-step inference
• Externalise, compress, and reconstruct work that exceeds the context window
• Derive an approach for a hard open problem with no standard playbook
• Measure the system honestly

Additional Responsibilities

• Work with decades of documents in a dozen languages with no schema
• Handle the same entity written five different ways
• Ensure the system is interrogable by a person accountable for a decision
• Build entirely on open-weight models without fine-tuning or external APIs
• Work within an air-gapped, self-hosted environment

Nice To Have

• Publication record
• Repository that does something nobody asked for

More Skills reasoning systems, retrieval, language models, multi-hop reasoning, temporal questions, decomposition, synthesis across sources, uncertainty propagation, architecture design, measurement

Other

• No fine-tuning
• No external APIs
• Self-hosted open-weight models
• Air-gapped deployment
• The models are fixed and architecture is the only lever
• No standard playbook exists
• Not this role: prompt templates, API integration, backend or UI, fine-tuning on labelled datasets
• They are reading for one thing: whether you can reason your way to an architecture, build it, and measure it honestly

Prepare for this role

Recommended resources to build the skills for this position. Sponsored.

Deep Learning Specialization

Coursera

Five-course deep learning series covering CNNs, RNNs, transformers, and ML strategy.

LangChain Chat with Your Data

Coursera

Build RAG applications with LangChain — document loading, splitting, embeddings, and retrieval.

Machine Learning Specialization

Coursera

Andrew Ng's updated ML course — regression, classification, neural networks, and decision trees.

More RAG jobs

Voice AI Intern

Atlys

Delhi

Today

AI Product Engineer

Cognologix Technologies Pvt. Ltd.

Pune

Today

Senior GenAI Engineer

FindQ

India

Today

GenAI Trainee

Larsen &

• Toubro

Powai

Today

Forward Deployed AI Engineer

Skit.ai

Bengaluru

Today

Software Engineer II, AI/ML

DigitalOcean

Bengalore

Today