Saurav Prateek

Software Engineer | Content Creator

I have a Bachelors degree in Computer Science Engineering (2015-19) and currently working as a Senior Web Solutions Engineer at Google. I have a keen interest in Machine Learning, Artificial Intelligence, Distributed Systems, System Design, Data Structures, Algorithms and Maths.

Get in touch GitHub LinkedIn

Experience

Senior Web Solutions Engineer (L5) at Google, India

2025 - Present

Working in the gTech Ads - TAI (Tools Automation and Infrastructure) team as an L5 individual contributor and Leading Headroom tool that enables Video Specialists across LCS and GCS to perform video headroom planning for their book of business using historical data from their advertiser.

Web Solutions Engineer II (L4) at Google, India

2022 - 2025

Working in the gTech-TDA (Tools Development and Automation) team as an L4 individual contributor and leading the development of multiple tools used by the Sales and Business team to automate and find insights in the gTech Ads ecosystem

Web Solutions Engineer I (L3) at Google, India

2021 - 2022

Worked in the gTech-TDA (Tools Development and Automation) team as an L3 Individual Contributor. Authored multiple RPCs to support the data import from the customer’s Google Ads Account. The data involved multiple Campaigns, Ad-Groups and other ads entities. Developed a recursion and back-tracking based solution to support the real-time generation of Product Groups (a google ads-entity) in bulk.

Software Engineer (SDE 1) at GeeksforGeeks

2019 - 2021

Worked on Amazon Dynamo DB built scalable architecture in No SQL design. Explored relational model access patterns and came up with a reasonable no sql design for the same. Learned to read through matrix and designed an architecture which could serve the query in millisecond order of time along with being cost efficient.

Personal Projects

Reproducing TinyStories small language models (SLM)

This repository contains a reproduction of the TinyStories language models described in the paper "TinyStories: How Small Can Language Models Be and Still Speak Coherent English?" by Ronen Eldan and Yuanzhi Li.

Python Jupyter Notebook Deep Neural Networks

Toy-Transformer (decoder-only)

A toy Transformer model (0.21M parameters)implementation from scratch! If you have a block of text on which you want to train a transformer model and let the model generate text afterwards, then you can use this repo to achieve all of it on your local machine without a GPU.

Python Deep Neural Networks

Deep Researcher: Sequential Plan Reflection and Candidates Crossover

Deep Researcher: Reflect Evolve

Deep Researcher is an advanced autonomous agent architecture designed to generate detailed, PhD-level research reports on complex topics. This system utilizes Sequential Research Plan Refinement and Candidates Crossover.

Python LangChain Google GenAI Deep Research

LangPost - AI Agent that creates a Linkedin Post

LangPost is an AI agent that creates a Linkedin post from a Linkedin newsletter article or any other blog post. It utilizes the concept of Few Shot encoding.

TypeScript LangChain LangGraph AI Engineering

Static Deep Research Agent (Static-DRA)

A configurable and hierarchical tree-based static Deep Research Agent designed to overcome the limitations of standard RAG pipelines for complex, multi-turn research tasks by utilizing a static workflow governed by two user-tunable parameters: Depth and Breadth.

Python LangChain Deep Research AI Engineering

Featured Talks

Design Patterns and Data Modeling in MongoDB

MongoDB Connect @ Microsoft, Noida • 2025

Optimizing Your Agentic Workflows with Parallel Execution

GDG Cloud Noida @ IIT Delhi • 2025

Presenting our work on 'AI Powered Request Workflows'

Hack-a-gTech @ Google Sunnyvale, California • 2025

Building and Deploying your Graph database using Neo4j

CNCG, Gurugram - Devtron Inc • 2024

Databases and related optimizations

GDG Cloud Noida @ IIT Delhi • 2024

Introduction to TensorFlow.js

Bennett University, Noida • 2024

Panel Discussion at Action AI Conference

Action AI Conference, New Delhi • 2024

YouTube

Visit Channel →

RAG and LangChain | Building multi-agent workflow in LangGraph

12 Episodes • 25K Views

Let's build a multi-agent workflow in LangGraph from scratch. We will build a RAG framework which takes an external knowledge base and grounds the LLM model. Further we will send some Human Prompts and see the RAG generating answers for the same.

AI Engineering 101

23 Episodes • 4K Views

All about AI Engineering, Agentic Workflows and Orchestration from scratch.

Distributed Systems 101

13 Episodes • 10K Views

A playlist on Distributed Systems where we will talk about concepts like Replication, Transactions, Databases internals, Partitioning and many more from scratch.