Ai / Machine Learning Engineer Llms & Rag Pipelines Bengaluru (India)

AI / Machine Learning Engineer (LLMs & RAG Pipelines)

Location: Bangalore & Gurgaon - Hybrid (India)

Employment Type: Full-Time

About Us

We are actively building the next generation of intelligent product ecosystems. Our team doesn't just talk about artificial intelligencewe deploy production-grade LLM pipelines, vector search workflows, and scalable AI infrastructure into real-world applications every single week. If you are passionate about generative AI, low-latency inference, and turning cutting-edge research into reliable software, you will find your home here.

The Role

We are seeking an creative AI / Machine Learning Engineer to take charge of our AI infrastructure. You will design, fine-tune, and scale artificial intelligence systems, integrating advanced LLM providers and vector databases into our core product architecture. You will work closely with our core engineering team to ensure our AI features are fast, cost-effective, and deeply integrated.

Key Responsibilities

RAG Implementation: Architect and scale Retrieval-Augmented Generation (RAG) frameworks using vector databases like Qdrant or Pinecone.

LLM & Prompt Optimization: Optimize prompt engineering workflows, manage token limits, and evaluate performance across various LLM providers (OpenAI, Anthropic, open-source models).

Inference Engineering: Build and maintain low-latency, scalable AI inference endpoints using Python and FastAPI.

Model Fine-Tuning: Fine-tune open-source models on domain-specific datasets to improve accuracy and business relevance.

Pipeline Monitoring: Monitor AI system accuracy, latency, and cost efficiency, implementing caching strategies and fallback mechanisms.

What We Are Looking For (Requirements)

Experience: 3+ years of professional software/ML engineering experience with a robust focus on production-grade AI systems.

Python Mastery: Deep expertise in Python, asynchronous programming, and up-to-date frameworks like FastAPI.

Vector & Database .