Smith . profile photo 1 on Tal

Smith .

Senior Associate Technology L1 • Delivery, Engineering

Gurugram

5+ years of work experience

Publicis Sapient
Oct 20241 yr 11 mosPresent

Gurugram, Haryana, India

AI Plan Diagnostic – Intelligent Chatbot System Designed and developed an AI-powered chatbot to respond to user queries based on real-time enterprise data. Architected the end-to-end system using Azure Web Apps, Azure OpenAI, and Databricks SQL Warehouse (serverless). Implemented real-time data retrieval from Databricks for dynamic response generation. Integrated agentic framework to orchestrate multi-step LLM reasoning and query processing. Used graph-based model embeddings to enhance LLM context with structured domain knowledge. Built APIs using FastAPI, with PostgreSQL for storing chat logs and history. Employed Docker for containerizing the entire application. Set up CI/CD pipelines using GitHub Actions for seamless deployment and integration. Selected and optimized LLMs via Azure OpenAI to improve response quality and system reliability.
Rupyz
Apr 20246 mosSep 2024

Gurugram, Haryana, India

PRODUCT DEVELOPMENT Engineered API endpoints to introduce a multicurrency price management system, empowering users to seamlessly manage product pricing in their preferred currency. This enhancement not only elevated user experience but also provided greater flexibility in pricing strategies, ultimately enriching product value and competitiveness. Spearheaded the development of API endpoints facilitating a streamlined multi-country login process, incorporating OTP verification via email and mobile numbers. By seamlessly integrating this feature and validating credentials against Django customer model, authentication token updates were made more secure and efficient, ensuring a frictionless user login experience while enhancing platform security. Designed and implemented robust APIs for real-time user activity tracking, capturing geo-location, longitude, and latitude alongside real-time activity and battery percentage information. This comprehensive tracking system not only enabled precise monitoring of user activities within the application but also provided invaluable insights into user behavior and usage patterns, driving informed decision-making and product optimization. Innovated and deployed APIs for bulk uploading of customer and product lists, supporting both CSV and Excel file formats and leveraging S3 storage for efficient data handling. This streamlined data management process significantly reduced manual effort and time required for data import, enhancing productivity and scalability for the organization. Pioneered the establishment of an automated deployment workflow through the creation of Makefile and Docker Compose files, effectively doubling deployment efficiency. By automating deployment processes, this initiative not only reduced the risk of human error but also accelerated the delivery of updates and enhancements to the product, ensuring rapid response to market demands and increasing overall operational efficiency. Developed API's with skills in
BDB.ai
Jan 20213 yrs 3 mosMar 2024

Bengaluru, Karnataka, India

Developed an ETL framework using PySpark for processing data in RDBMS databases and MongoDB and apply any script on top of that for analysis. With schedulable and chainable jobs, the framework accelerated use case development by 3x, enhancing productivity and automating deployment processes. Created the backend for "Datastore Load," optimizing data storage for efficient business storytelling. Streamlining data acquisition with user queries or sandbox file processing, the system establishes tailored database connections, inserts data into tables, and generates synonyms critical for NLP. Upgraded Python version from 3.6.5 to 3.10.11 across multiple projects, encompassing dependencies such as MongoDB, Confluent Kafka, RDBMS databases, etc. By adjusting specific version of libraries, achieved 2x to 3x efficiency improvements, resolved existing bugs, and eliminated vulnerabilities. Devised a monitoring solution for data production, consumption, rates, and sizes, enabling platform users to analyse data for informed decision-making, resulting in a 20% increase in platform usability through seamless integration with the existing data infrastructure. Created a testing mechanism for data pipelines by generating multiple test cases with input/expected data and schema, enabling customers to develop error-free pipelines and reducing effort by 50% in pipeline construction. Created and encouraged innovation by developing external offset management for Kafka, utilizing MongoDB to store processed offsets, ensuring 2-fold increased resiliency and fault tolerance in handling failures and restarts.
GitHub
@SmithSaurav

Smith’s AI Stack

claude code
claude code
codex
codex
gemini
gemini
github copilot
github copilot
langchain
langchain
langgraph
langgraph

Education

nitte meenakshi institute of technology

Bachelor of Engineering (B.E.), Computer Science

nitte meenakshi institute of technology

20124 yrs2016

11th &12th, PCM

R K Public School

20102 yrs2012

Join Smith and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Smith . on Tal

Headline

Senior Associate Technology L1 • Delivery, Engineering

About

I’m a backend and AI engineer passionate about building intelligent, scalable, and high-performance systems that bridge data, AI, and product innovation. Currently at Publicis Sapient, I design and develop AI-driven chatbot systems that leverage Azure OpenAI, Databricks, and FastAPI to deliver dynamic, context-aware enterprise insights in real time. My work involves orchestrating multi-agent reasoning frameworks, optimizing LLM embeddings, and deploying fully containerized applications with Docker and CI/CD pipelines on Azure. Before this, at Rupyz Fintech, I built and optimized backend services for a fast-growing fintech product — from multi-currency pricing APIs and OTP-based multi-country login systems to real-time user activity tracking and automated deployments. At Bizviz Technology, I engineered scalable ETL frameworks in PySpark, built data monitoring systems, and containerized microservices with Django and Kubernetes, enabling teams to accelerate analytics and ML workflows by 3–4x. With hands-on expertise in Python, Django, FastAPI, SQL, MongoDB, Azure, AWS, and Docker, I enjoy solving complex data challenges and designing systems that transform raw data into intelligent experiences. I’m currently pursuing my MS in AI & ML from Scaler Neovarsity while continuously exploring ways to integrate AI agents, data engineering, and cloud technologies into real-world products. 🚀 Always open to discussions around AI systems, backend architecture, or anything that involves turning data into intelligence.

Location

Gurugram

AI stack

claude code

codex

gemini

github copilot

langchain

langgraph

Work experience

Senior Associate Technology L1 • Delivery, Engineering at Publicis Sapient, Oct 2024 - Present · 1 yr 11 mos, Gurugram, Haryana, India, AI Plan Diagnostic – Intelligent Chatbot System Designed and developed an AI-powered chatbot to respond to user queries based on real-time enterprise data. Architected the end-to-end system using Azure Web Apps, Azure OpenAI, and Databricks SQL Warehouse (serverless). Implemented real-time data retrieval from Databricks for dynamic response generation. Integrated agentic framework to orchestrate multi-step LLM reasoning and query processing. Used graph-based model embeddings to enhance LLM context with structured domain knowledge. Built APIs using FastAPI, with PostgreSQL for storing chat logs and history. Employed Docker for containerizing the entire application. Set up CI/CD pipelines using GitHub Actions for seamless deployment and integration. Selected and optimized LLMs via Azure OpenAI to improve response quality and system reliability.

Senior Software Engineer at Rupyz, Apr 2024 - Sep 2024 · 6 mos, Gurugram, Haryana, India, PRODUCT DEVELOPMENT Engineered API endpoints to introduce a multicurrency price management system, empowering users to seamlessly manage product pricing in their preferred currency. This enhancement not only elevated user experience but also provided greater flexibility in pricing strategies, ultimately enriching product value and competitiveness. Spearheaded the development of API endpoints facilitating a streamlined multi-country login process, incorporating OTP verification via email and mobile numbers. By seamlessly integrating this feature and validating credentials against Django customer model, authentication token updates were made more secure and efficient, ensuring a frictionless user login experience while enhancing platform security. Designed and implemented robust APIs for real-time user activity tracking, capturing geo-location, longitude, and latitude alongside real-time activity and battery percentage information. This comprehensive tracking system not only enabled precise monitoring of user activities within the application but also provided invaluable insights into user behavior and usage patterns, driving informed decision-making and product optimization. Innovated and deployed APIs for bulk uploading of customer and product lists, supporting both CSV and Excel file formats and leveraging S3 storage for efficient data handling. This streamlined data management process significantly reduced manual effort and time required for data import, enhancing productivity and scalability for the organization. Pioneered the establishment of an automated deployment workflow through the creation of Makefile and Docker Compose files, effectively doubling deployment efficiency. By automating deployment processes, this initiative not only reduced the risk of human error but also accelerated the delivery of updates and enhancements to the product, ensuring rapid response to market demands and increasing overall operational efficiency. Developed API's with skills in

Software Engineer at BDB.ai, Jan 2021 - Mar 2024 · 3 yrs 3 mos, Bengaluru, Karnataka, India, Developed an ETL framework using PySpark for processing data in RDBMS databases and MongoDB and apply any script on top of that for analysis. With schedulable and chainable jobs, the framework accelerated use case development by 3x, enhancing productivity and automating deployment processes. Created the backend for "Datastore Load," optimizing data storage for efficient business storytelling. Streamlining data acquisition with user queries or sandbox file processing, the system establishes tailored database connections, inserts data into tables, and generates synonyms critical for NLP. Upgraded Python version from 3.6.5 to 3.10.11 across multiple projects, encompassing dependencies such as MongoDB, Confluent Kafka, RDBMS databases, etc. By adjusting specific version of libraries, achieved 2x to 3x efficiency improvements, resolved existing bugs, and eliminated vulnerabilities. Devised a monitoring solution for data production, consumption, rates, and sizes, enabling platform users to analyse data for informed decision-making, resulting in a 20% increase in platform usability through seamless integration with the existing data infrastructure. Created a testing mechanism for data pipelines by generating multiple test cases with input/expected data and schema, enabling customers to develop error-free pipelines and reducing effort by 50% in pipeline construction. Created and encouraged innovation by developing external offset management for Kafka, utilizing MongoDB to store processed offsets, ensuring 2-fold increased resiliency and fault tolerance in handling failures and restarts.

Education

Bachelor of Engineering (B.E.), Computer Science, nitte meenakshi institute of technology

11th &12th, PCM, R K Public School

Public links

github: https://github.com/SmithSaurav