Data Scientist

Role: Data Scientist

Exp: 3 to 5+ Years

Location : Hyderabad

About the Role

We are looking for a sharp, hands-on AI Engineer experience in Large Language Models (LLMs) and Agentic AI to work alongside a Senior AI Scientist. You will contribute to the design, fine-tuning, and deployment of AI models, enabling goal-driven, autonomous AI agents to perform complex reasoning and decision-making tasks. Your role will involve building and optimizing LLMs, integrating them with retrieval-augmented generation (RAG), and enhancing their efficiency for real-time automation.

Key Responsibilities:

1. Agentic AI Development & Workflow Optimization

2. Develop autonomous AI agents capable of dynamic decision-making and task execution.

3. Work with LangGraph, LangChain, LlamaIndex, and other frameworks to create multi-agent workflows.
4. Optimize goal-driven AI behavior for handling complex automation and reasoning tasks.

5. Enhance agent collaboration mechanisms, improving their ability to break down goals, delegate tasks, and refine execution strategies.

6. Language Model Fine-Tuning & Optimization

7. Work with Tiny LLMs / Small LLMs (e.g., Phi-3, OpenELM, Mistral, Llama 3) to build efficient, scalable AI models.

8. Fine-tune models using full fine-tuning, LoRA, QLoRA, and PEFT techniques to improve task-specific performance.
9. Optimize GPU memory and compute resources to efficiently run fine-tuned models for real-world applications.

10. Research and implement efficient inference strategies for deploying LLMs in production.

11. Retrieval-Augmented Generation (RAG) & Knowledge Integration

12. Design and implement RAG architectures to improve contextual awareness and decision-making capabilities of AI agents.

13. Work with vector databases (e.g., ChromaDB, FAISS, Weaviate) for efficient knowledge retrieval and document understanding.

14. Develop techniques for continuous learning, enabling agents to refine knowledge over time.

15. NLP & Intelligent Document Processing

16. Work on OCR-based automation, enhancing document understanding using Tesseract, PaddleOCR, and AI-powered text extraction.

17. Develop custom tokenization, embeddings, and NLP pipelines for document classification, summarization, and intent recognition.

18. Implement domain-specific adaptations of LLMs to improve accuracy and performance in structured and unstructured text processing.

19. AI Infrastructure & Compute Optimization

20. Configure fine-tuning infrastructure, manage GPU/TPU memory, and optimize compute for LLM fine-tuning.

21. Experiment with different fine-tuning strategies to optimize models for low-latency, high-performance applications.

Required Skills & Qualifications Core AI & NLP Expertise

• Strong background in NLP, LLMs, and deep learning with experience in fine-tuning and optimizing models.
• Proficiency in Python and experience with Hugging Face Transformers, LangChain, LlamaIndex, and ML frameworks (PyTorch, TensorFlow, scikit-learn).
• Experience with Tiny LLMs / Edge AI and optimizing Phi-3, OpenELM, and Mistral for efficiency.
• RAG & Data EngineeringSolid understanding of Retrieval-Augmented Generation (RAG) and its application in agentic AI workflows.
• Experience working with SQL & NoSQL databases (PostgreSQL, MongoDB, ChromaDB, FAISS, etc.).
• Familiarity with big data processing frameworks like Spark, Dask, or Ray for handling large-scale workloads.

AI Compute & Infrastructure (Plus, Not Mandatory)

• Experience deploying AI models using Docker, Kubernetes, or serverless architectures.
• Knowledge of MLOps best practices for model retraining, monitoring, and optimization.

Why Join Us?

• Work on cutting-edge AI-powered agentic automation at the intersection of LLMs, RAG, and multi-agent systems.
• Be part of a team building goal-driven, autonomous AI agents that redefine how businesses leverage AI.
• Collaborate with a highly skilled Senior AI Scientist and grow your expertise in LLM fine-tuning, optimization, and AI-driven automation.
• Competitive compensation and access to state-of-the-art AI models and compute resources.