Virat Srivastava profile photo 2 on Tal
Virat Srivastava profile photo 1 on Tal
Virat Srivastava profile photo 2 on Tal
Virat Srivastava profile photo 1 on Tal

Virat Srivastava

Software Architect @Starside Technologies | MTech AI

India

3+ years of work experience

Starside Technologies
Jan 202610 mosPresent

Bengaluru

Lead the design and development of AI-powered enterprise software platforms. - Architect backend systems using Python, Go, PostgreSQL, and AWS cloud infrastructure. - Develop workflow automation, compliance management, and document-processing solutions for aerospace and engineering organizations. - Design and implement AI/ML pipelines, report-generation systems, and intelligent decision-support tools. - Define system architecture, microservices, APIs, data models, and security standards. - Manage software development lifecycle, including requirements analysis, planning, implementation, testing, and deployment. -Supervise and mentor junior engineers and coordinate technical activities across development teams. - Oversee cloud infrastructure, DevOps processes, scalability planning, and production operations. - Collaborate with stakeholders to translate business and regulatory requirements into technical solutions. - Lead research and innovation initiatives involving artificial intelligence, automation, compliance monitoring, and workflow orchestration.
FirstEigen
Aug 20256 mosJan 2026

Illinois, United States

Introduced Agentic and Python Workflows in Databuck's BuckGPT. - Designed and Develop Agentic Pipelines for Integrating with system Data for providing insights. - Designed and implemented Chatbot API key integration, enabling secure and scalable access to LLM services. - Developed Buck GPT v1 scheduler using LangChain and LangGraph, enabling automated and time-based orchestration of AI workflows. - Designed and implemented the AI Summary generation feature, followed by enhancements including database persistence, robust error handling, and regeneration logic. - Built the Range Summary feature with database support, including intelligent reuse vs regeneration logic based on summary generation timestamps. - Designed and implemented Buck GPT v2 Intent Classification Pipeline, comprising 12 modular pipelines with: - Variable processing logic per intent - Database-backed state management - Embedding generation using three different embedding models - Implemented Entity Classification as an extension to the intent classification framework to enable structured extraction and downstream rule-based actions. - Conducted a comparative study of multiple embedding models, evaluating performance, accuracy, and suitability across different use cases. - Architected and developed a Redis–PostgreSQL hot–cold state system for intent classification: - Redis for low-latency hot state - PostgreSQL for cold state archival - Modular intent + entity classifier with rule-book driven actions - Full database logging of actions and decisions in a predefined schema - Researched and devised multiple strategies for managing context window overflow in large-volume range summary generation, ensuring scalability without loss of semantic accuracy. - Performed SMTP/IMAP protocol research and developed a mock implementation of an Email Agent for automated email ingestion and processing. - Developed and maintained the documentation for all developed features
Vellore Institute of Technology
May 202511 mosMar 2026

Bhopal

- Develop end-to-end retrieval-augmented generation pipelines (LangChain + Pinecone) to fuse vector-based document retrieval with LLM outputs, boosting answer relevance. - Architect agentic AI frameworks that orchestrate multi-step “thought–action” loops, integrating custom tools (APIs, DB queries) for autonomous decision-making workflows. - Drive master’s thesis research under faculty guidance—benchmarking transformer models, defining novel evaluation metrics for context-aware agents, and publishing experimental insights.
AntiProton Labs
Oct 20231 yr 9 mosJun 2025

Bhopal

At AntiProton Labs, I co-founded a startup aimed to develop Ai based cargo transport through autonomous gait planning and hexapod development. Utilizing ROS2 and NVIDIA Isaac Sim, we made significant strides in robotics and AI-powered mobility solutions.
GitHub
@ViratSrivastava

Virat’s AI Stack

anthropic apis
anthropic apis
antigravity
antigravity
claude code
claude code
claude
claude
gemini
gemini
langgraph
langgraph

Education

Vellore Institute of Technology

Master of Technology - MTech, Artificial Intelligence

Vellore Institute of Technology

20215 yrs2026
Campion school bhopal

Higher Secondary Certificate (HSC), Science, Mathematics and Computer Science

Campion school bhopal

200714 yrs2021

Join Virat and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Virat Srivastava on Tal

Headline

Software Architect @Starside Technologies | MTech AI

About

Building StarSide Technologies core product - Assemble It MTech in AI at Vellore Institute of Technology

Location

India

AI stack

anthropic apis

antigravity

claude code

claude

gemini

langgraph

Work experience

Software Architect at Starside Technologies, Jan 2026 - Present · 10 mos, Bengaluru, Lead the design and development of AI-powered enterprise software platforms. - Architect backend systems using Python, Go, PostgreSQL, and AWS cloud infrastructure. - Develop workflow automation, compliance management, and document-processing solutions for aerospace and engineering organizations. - Design and implement AI/ML pipelines, report-generation systems, and intelligent decision-support tools. - Define system architecture, microservices, APIs, data models, and security standards. - Manage software development lifecycle, including requirements analysis, planning, implementation, testing, and deployment. -Supervise and mentor junior engineers and coordinate technical activities across development teams. - Oversee cloud infrastructure, DevOps processes, scalability planning, and production operations. - Collaborate with stakeholders to translate business and regulatory requirements into technical solutions. - Lead research and innovation initiatives involving artificial intelligence, automation, compliance monitoring, and workflow orchestration.

Agentic AI Engineer at FirstEigen, Aug 2025 - Jan 2026 · 6 mos, Illinois, United States, Introduced Agentic and Python Workflows in Databuck's BuckGPT. - Designed and Develop Agentic Pipelines for Integrating with system Data for providing insights. - Designed and implemented Chatbot API key integration, enabling secure and scalable access to LLM services. - Developed Buck GPT v1 scheduler using LangChain and LangGraph, enabling automated and time-based orchestration of AI workflows. - Designed and implemented the AI Summary generation feature, followed by enhancements including database persistence, robust error handling, and regeneration logic. - Built the Range Summary feature with database support, including intelligent reuse vs regeneration logic based on summary generation timestamps. - Designed and implemented Buck GPT v2 Intent Classification Pipeline, comprising 12 modular pipelines with: - Variable processing logic per intent - Database-backed state management - Embedding generation using three different embedding models - Implemented Entity Classification as an extension to the intent classification framework to enable structured extraction and downstream rule-based actions. - Conducted a comparative study of multiple embedding models, evaluating performance, accuracy, and suitability across different use cases. - Architected and developed a Redis–PostgreSQL hot–cold state system for intent classification: - Redis for low-latency hot state - PostgreSQL for cold state archival - Modular intent + entity classifier with rule-book driven actions - Full database logging of actions and decisions in a predefined schema - Researched and devised multiple strategies for managing context window overflow in large-volume range summary generation, ensuring scalability without loss of semantic accuracy. - Performed SMTP/IMAP protocol research and developed a mock implementation of an Email Agent for automated email ingestion and processing. - Developed and maintained the documentation for all developed features

Research Assistant at Vellore Institute of Technology, May 2025 - Mar 2026 · 11 mos, Bhopal, - Develop end-to-end retrieval-augmented generation pipelines (LangChain + Pinecone) to fuse vector-based document retrieval with LLM outputs, boosting answer relevance. - Architect agentic AI frameworks that orchestrate multi-step “thought–action” loops, integrating custom tools (APIs, DB queries) for autonomous decision-making workflows. - Drive master’s thesis research under faculty guidance—benchmarking transformer models, defining novel evaluation metrics for context-aware agents, and publishing experimental insights.

Co-Founder at AntiProton Labs, Oct 2023 - Jun 2025 · 1 yr 9 mos, Bhopal, At AntiProton Labs, I co-founded a startup aimed to develop Ai based cargo transport through autonomous gait planning and hexapod development. Utilizing ROS2 and NVIDIA Isaac Sim, we made significant strides in robotics and AI-powered mobility solutions.

Education

Master of Technology - MTech, Artificial Intelligence, Vellore Institute of Technology

Higher Secondary Certificate (HSC), Science, Mathematics and Computer Science, Campion school bhopal

Public links

github: https://github.com/ViratSrivastava