Vivek Singh profile photo 4 on Tal
Vivek Singh profile photo 1 on Tal
Vivek Singh profile photo 2 on Tal
Vivek Singh profile photo 3 on Tal
Vivek Singh profile photo 4 on Tal
Vivek Singh profile photo 1 on Tal

Vivek Singh

Software Engineer @ VIP | Applied AI & ML Systems Engineer | Distributed Systems | High-Performance Pipelines

Mumbai

1+ year of work experience

VIP (Vermont Information Processing)
Jun 20251 yr 5 mosPresent

Mumbai, Maharashtra, India

A graduate of the Xavier Institute of Engineering with a Bachelor of Engineering (BE), currently contributing as an Artificial Intelligence Engineer at Vermont Information Processing (VIP). Core competencies include developing AI-driven solutions for complex business processes, optimizing large-scale systems, and implementing business intelligence frameworks to enhance decision-making within the beverage industry. At VIP, supported the creation of RGM_COPILOT, an AI agent server that streamlined pricing workflows and reduced response times through innovative intent routing. Contributed to RGM_ANALYTICS, a Natural Language Business Intelligence Engine, by engineering a system to enable data-driven insights. Passionate about leveraging serverless frameworks and high-availability clustering to drive operational efficiency, while fostering collaboration and innovation within cross-functional teams.
Jun 20251 yr 5 mosPresent

Project: RGM_COPILOT (Enterprise Multi-Agent Orchestrator) • Built a multi-tenant AI agent server deployed via Docker, using FastAPI and LangGraph to automate complex pricing workflows for the beverage industry. • Cut LLM API costs and reduced response time by designing a 3-layer intent routing system, which successfully bypasses the LLM and uses deterministic logic for ∼80% of routine user queries. • Refactored a monolithic 3,338-line Python script into 55 clean, testable modules, significantly improving system stability and allowing safe integration of 12+ specialized pricing tools. Project: RGM_ANALYTICS (Natural Language Business Intelligence Engine) • Engineered a Natural Language-to-ElasticSearch engine that translates conversational questions into safe database queries across a massive 209-field commercial dataset. • Unlocked net-new business value for the client by enabling the AI to correlate complex data points, discovering and generating high-value custom reports (like micro-market margin leakages) that were previously inaccessible. • Protected the live ElasticSearch 6.5 cluster by building a validation proxy layer that automatically blocks heavy, nested AI-generated queries from crashing the database. Project: Info-AI bridge (Secure Enterprise Search & RAG Platform) • Developed a secure enterprise search (RAG) platform that pulls data from Jira, Confluence, and Zendesk using automated BullMQ and Redis pipelines. • Solved critical enterprise data privacy requirements by mapping user Okta permissions directly into vector search filters, ensuring users can never retrieve or view restricted documents.
Jun 20251 yr 5 mosPresent

Project: RGM_COPILOT (Enterprise Multi-Agent Orchestrator) • Constructed a multi-tenant AI agent server deployed via Docker, using FastAPI and LangGraph to automate complex pricing workflows for the beverage industry. • Reduced inference latency by ∼850ms and cut LLM API compute costs by 75% by architecting a 3-layer semantic routing engine, successfully diverting 80% of routine traffic to deterministic fast-paths. • Architected a highly extensible multi-agent orchestration framework, refactoring a legacy 3,300+ line monolith into a modular, unit-tested system using memory profiling and async execution, accelerating the integration of 12+ external tooling APIs by 50%. Project: RGM_ANALYTICS (Natural Language Business Intelligence Engine) • Designed a Natural Language-to-ElasticSearch engine that translates conversational questions into safe database queries across a massive 209-field commercial dataset. • Unlocked net-new business value for the client by enabling the AI to correlate complex data points, discovering and generating high-value custom reports (like micro-market margin leakages) that were previously inaccessible. • Engineered a fault-tolerant validation proxy for a live ElasticSearch cluster handling 150+ peak QPS, preventing out-of-memory crashes by intercepting and shedding load from unoptimized, nested AI-generated queries. Project: Info-AI bridge (Secure Enterprise Search & RAG Platform) • Developed a secure enterprise search (RAG) platform that pulls data from Jira, Confluence, and Zendesk using automated BullMQ and Redis pipelines. • Solved critical enterprise data privacy requirements by mapping user Okta permissions directly into vector search filters, ensuring users can never retrieve or view restricted documents.
Bhabha Atomic Research Centre
Jun 20242 mosJul 2024

• Overview: Developed a scalable Content Management System (CMS) for the Saraswati website’s “SIR Bulletin” section, incorporating role-based access control and secure user management, supporting over 50+ users. • Contribution: Engineered a backend infrastructure using PHP, Apache, and MySQL on a XAMPP server, optimizing performance to handle 100+ daily user interactions, with jQuery enhancing data interactions and integration within the CMS. • Impact: Enhanced the content management workflow, contributing to a 35% increase in user engagement and overall satisfaction.
GitHub
@Vivek-1102

Vivek’s AI Stack

claude code
claude code
claude
claude
gemini
gemini
hugging face
hugging face
langgraph
langgraph
langsmith
langsmith

Education

Xavier Institute of Engineering

Bachelor of Engineering - BE

Xavier Institute of Engineering

20214 yrs2025

High School Diploma

Kirti Diam

Join Vivek and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Vivek Singh on Tal

Headline

Software Engineer @ VIP | Applied AI & ML Systems Engineer | Distributed Systems | High-Performance Pipelines

Location

Mumbai

AI stack

claude code

claude

gemini

hugging face

langgraph

langsmith

Work experience

Software Trainee | AI Engineer at VIP (Vermont Information Processing), Jun 2025 - Present · 1 yr 5 mos, Mumbai, Maharashtra, India, A graduate of the Xavier Institute of Engineering with a Bachelor of Engineering (BE), currently contributing as an Artificial Intelligence Engineer at Vermont Information Processing (VIP). Core competencies include developing AI-driven solutions for complex business processes, optimizing large-scale systems, and implementing business intelligence frameworks to enhance decision-making within the beverage industry. At VIP, supported the creation of RGM_COPILOT, an AI agent server that streamlined pricing workflows and reduced response times through innovative intent routing. Contributed to RGM_ANALYTICS, a Natural Language Business Intelligence Engine, by engineering a system to enable data-driven insights. Passionate about leveraging serverless frameworks and high-availability clustering to drive operational efficiency, while fostering collaboration and innovation within cross-functional teams.

Artificial Intelligence Engineer at Vermont Information Processing (VIP), Jun 2025 - Present · 1 yr 5 mos, Project: RGM_COPILOT (Enterprise Multi-Agent Orchestrator) • Built a multi-tenant AI agent server deployed via Docker, using FastAPI and LangGraph to automate complex pricing workflows for the beverage industry. • Cut LLM API costs and reduced response time by designing a 3-layer intent routing system, which successfully bypasses the LLM and uses deterministic logic for ∼80% of routine user queries. • Refactored a monolithic 3,338-line Python script into 55 clean, testable modules, significantly improving system stability and allowing safe integration of 12+ specialized pricing tools. Project: RGM_ANALYTICS (Natural Language Business Intelligence Engine) • Engineered a Natural Language-to-ElasticSearch engine that translates conversational questions into safe database queries across a massive 209-field commercial dataset. • Unlocked net-new business value for the client by enabling the AI to correlate complex data points, discovering and generating high-value custom reports (like micro-market margin leakages) that were previously inaccessible. • Protected the live ElasticSearch 6.5 cluster by building a validation proxy layer that automatically blocks heavy, nested AI-generated queries from crashing the database. Project: Info-AI bridge (Secure Enterprise Search & RAG Platform) • Developed a secure enterprise search (RAG) platform that pulls data from Jira, Confluence, and Zendesk using automated BullMQ and Redis pipelines. • Solved critical enterprise data privacy requirements by mapping user Okta permissions directly into vector search filters, ensuring users can never retrieve or view restricted documents.

Software Engineer at Vermont Information Processing (VIP), Jun 2025 - Present · 1 yr 5 mos, Project: RGM_COPILOT (Enterprise Multi-Agent Orchestrator) • Constructed a multi-tenant AI agent server deployed via Docker, using FastAPI and LangGraph to automate complex pricing workflows for the beverage industry. • Reduced inference latency by ∼850ms and cut LLM API compute costs by 75% by architecting a 3-layer semantic routing engine, successfully diverting 80% of routine traffic to deterministic fast-paths. • Architected a highly extensible multi-agent orchestration framework, refactoring a legacy 3,300+ line monolith into a modular, unit-tested system using memory profiling and async execution, accelerating the integration of 12+ external tooling APIs by 50%. Project: RGM_ANALYTICS (Natural Language Business Intelligence Engine) • Designed a Natural Language-to-ElasticSearch engine that translates conversational questions into safe database queries across a massive 209-field commercial dataset. • Unlocked net-new business value for the client by enabling the AI to correlate complex data points, discovering and generating high-value custom reports (like micro-market margin leakages) that were previously inaccessible. • Engineered a fault-tolerant validation proxy for a live ElasticSearch cluster handling 150+ peak QPS, preventing out-of-memory crashes by intercepting and shedding load from unoptimized, nested AI-generated queries. Project: Info-AI bridge (Secure Enterprise Search & RAG Platform) • Developed a secure enterprise search (RAG) platform that pulls data from Jira, Confluence, and Zendesk using automated BullMQ and Redis pipelines. • Solved critical enterprise data privacy requirements by mapping user Okta permissions directly into vector search filters, ensuring users can never retrieve or view restricted documents.

Project Trainee at Bhabha Atomic Research Centre, Jun 2024 - Jul 2024 · 2 mos, • Overview: Developed a scalable Content Management System (CMS) for the Saraswati website’s “SIR Bulletin” section, incorporating role-based access control and secure user management, supporting over 50+ users. • Contribution: Engineered a backend infrastructure using PHP, Apache, and MySQL on a XAMPP server, optimizing performance to handle 100+ daily user interactions, with jQuery enhancing data interactions and integration within the CMS. • Impact: Enhanced the content management workflow, contributing to a 35% increase in user engagement and overall satisfaction.

Education

Bachelor of Engineering - BE, Xavier Institute of Engineering

High School Diploma, Kirti Diam

Public links

github: https://github.com/Vivek-1102