Job title : AI/ML Research Scientist

Exp : 18 to 23years

Location : Bangalore

Key Responsibilities :

- Drive research and development in LLMs, Vision Transformers (ViTs), Diffusion Models, Reinforcement Learning (RL), and Multi-Agent Systems

- Architect and implement RAG-based knowledge systems using vector databases like Azure AI Search, Databricks, or similar

- Fine-tune LLMs using techniques like LoRA / QLoRA for domain-specific applications

- Design and develop real-time RAG systems for dynamic, context-aware decision making

- Utilize graph-based RAG techniques and Graph Neural Networks (GNNs) for enhanced contextual reasoning

- Integrate multimodal transformers combining text, image, and audio data

- Lead performance optimization efforts using Redis caching, semantic indexing, and latency reduction techniques

Key Skills :

- GenAI, Python coding

- Research-level knowledge of LLMs, vision transformers (ViTs), diffusion models, RL, or multi-agent systems

- RAG & Vector Databases: Expertise in building and querying knowledge bases using Retrieval-Augmented Generation with Azure AI Search or similar technologies

- LLM Fine-Tuning: Hands-on experience with efficient finetuning techniques (LoRA/QLoRA) for specializing models on custom datasets

- Proficiency with libraries for data transformation and comparison, such as JSON Patch and DeepDiff

- Quantum Computing: Understanding of quantum algorithms and tools like IBM's Qiskit and Google's Cirq.

- Familiarity with Multimodal transformers - integrating text, image, and audio data to create models

- Experience on graph-based RAGs for contextual reasoning and incorporating knowledge connections from graph neural networks. Real time RAG systems to handle dynamic and up-to-date information

- Track record of research (papers, patents, open source)

Must have skills :

- LLM & RAG Architecture Expertise: (Must have for 30 and 29 differentiating factor will be level of expertise)

- Hands-on experience with Retrieval-Augmented Generation (RAG) architectures using embeddings via Azure AI Search and Databricks.

- Proficient in implementing semantic search capabilities.

- Familiar with MCP servers for scalable deployment.

- Agentic AI & Orchestration: (Must have for 30 and 29 differentiating factor will be level of expertise)

- Experience with autonomous decision support using LangGraph.

- Skilled in agent orchestration using Microsoft Copilot Studio and CrewAI.

- Performance Optimization: (Must have for 30 and 29 differentiating factor will be level of expertise)

- Working knowledge of latency reduction techniques for RAG-based applications, including Redis-based caching.

- LLM Fine-Tuning: (Must have for 30 and 29 differentiating factor will be level of expertise)

- Practical understanding of fine-tuning methods such as LoRA (Low-Rank Adaptation).

- Model Selection & Prompting: (Must have for 30 and 29 differentiating factor will be level of expertise)

- Awareness of the latest LLMs tailored to specific use cases (e.g., Claude, Gemini, GPT series).

- Understanding of prompt engineering requirements across different models.

- Cost Estimation: (Must have for 30 and 29 differentiating factor will be level of expertise)

- Ability to calculate and optimize costs for API-based model usage.

- Functional/Team experience (Must have for 29 and 30)

- Expertise with diverse AI uses cases - Must for Grade 30 and 29

- Business and Domain Understanding : Must for Grade 30 and 29

- Track record of research (papers, patents, open source)

- Client management for SG31 & Sg30