Sanathraj Narayan

Sanathraj Narayan

Data Science Manager

About

Sanath brings over a decade of experience in AI/ML, data science, and analytics, with a strong track record of building and deploying machine learning solutions at scale. Before joining Lam, he was a Senior Data Scientist at Ericsson, where he led model development and implementation for network rollout and forecasting use cases. He has also worked at Mindtree and KPMG, focusing on predictive analytics, scalable ML models, and enterprise AI solutions. Sanath is passionate about industrializing AI/ML models and driving real-world impact. He has been an active speaker at AI/ML conferences like Cypher and DataHack Summit, sharing insights on LangChain and LLM-based applications.

This session provides a hands-on, engineering-focused comparison of Large Language Models (LLMs) and Small Language Models (SLMs) in real-world applications. Participants will see both models implemented side by side across two key paradigms: Retrieval-Augmented Generation (RAG) and agentic workflows. We start by building a RAG pipeline with an LLM and replicate it using an SLM, comparing performance across quality, latency, cost, and consistency. The session then extends to a simple multi-agent workflow (planner–executor), evaluating both approaches on reasoning, tool usage, and robustness, along with the impact of optimizations like prompt design, fine-tuning, and memory. By the end, participants will gain a practical framework for choosing between LLMs and SLMs based on use case, constraints, and scale. 

Read More →

This full-day, hands-on workshop equips participants to design, build, and optimize real-world AI systems powered by Small Language Models (SLMs). Unlike traditional LLM-heavy approaches, this workshop focuses on cost-efficient, production-aware architectures that run entirely within Google Colab’s free tier—making advanced AI engineering accessible without expensive compute infrastructure. 

Participants will progress through six tightly integrated modules, building toward a complete multi-agent, RAG-enabled AI system that solves a real-world problem. Every module includes end-to-end hands-on demos with pre-configured notebooks that participants take home after the session. 

Key Learning Outcomes 

By the end of this workshop, participants will be able to: 

  • Deploy and run inference with SLMs (Phi-3 Mini, Gemma 2B, TinyLlama) within Google Colab free-tier limits 
  • Apply quantization techniques  and parameter-efficient fine-tuning with QLoRA for domain-specific tasks 
  • Build a lightweight RAG pipeline with vector search, connecting fine-tuned SLMs to external knowledge bases 
  • Design and orchestrate multi-agent workflows using role-specialized SLMs with shared state and minimal memory overhead 
  • Architect an end-to-end Agentic RAG system combining retrieval, reasoning, and generation under constrained compute 
  • Simulate edge deployment using llama.cpp with GGUF models, profiling latency and optimizing for CPU-first execution 
Read More →
View All Speakers

We use cookies essential for this site to function well. Please click to help us improve its usefulness with additional cookies. Learn about our use of cookies in our Privacy Policy & Cookies Policy.

Show details