Suyash Sanjay profile photo 2 on Tal
Suyash Sanjay profile photo 1 on Tal
Suyash Sanjay profile photo 2 on Tal
Suyash Sanjay profile photo 1 on Tal

Suyash Sanjay

Founding Engineer @ Anko AI | Building LLM Agents, MCP Integrations & Multi-Agent Systems that Scale

Bengaluru

2+ years of work experience

Anko AI
Jul 20242 yrs 2 mosPresent

Bengaluru

- Engineered a human-in-the-loop memory pipeline: an LLM agent extracts structured memory candidates from historical chat sessions, validates them against a live entity registry, and promotes approved facts into future analyst prompts — enabling continuous agent learning. - Designed and built end-to-end MCP integration enabling secure external client access to a production chat platform, including session persistence, token-based authentication, and cross-service request routing. - Built a production LangGraph functional FSM for multi-turn LLM agent execution with SSE streaming, tool validation, reasoning/response separation, and runaway-loop safeguards. - Designed a hierarchical multi-agent system where sub-agents are invoked as structured tools from a parent Analyst agent, enabling compositional analytics reasoning. - Implemented LLM-driven code execution tools allowing agents to dynamically write and run Python/pandas scripts and generate Word reports within the chat loop. - Built an internal prompt engineering platform (Eva) with versioned Jinja2 templates, auto-discovered MKG variable injection, and a dynamic skills system that injects new prompts and tools into live agent executions at runtime. - Migrated the ThoughtSpot integration to the official REST API SDK with stateless bearer token auth and automatic 401 retry, providing a reliable data query layer for LLM agents at runtime. - Designed and architected the frontend with a service layer and modular design, reducing build times by 50% and producing a codebase structure that LLMs can navigate and extend reliably.
Anko AI
Jan 20247 mosJul 2024

Bengaluru, Karnataka, India

• Spearheaded the development of a new frontend service for the MVP, which resulted in the company's first paid customer. • Designed and built an admin panel for metric setup and visualization using Vis.js, reducing manual setup time by 30%. • Upgraded the frontend service from JavaScript to TypeScript React, improving type safety, maintainability, and scalability. • Implemented a Snowflake integration panel to create and update data models, cutting data integration time by 70%.
Rachnaye
Jun 20232 mosJul 2023

Bengaluru, Karnataka, India

LeetCode
@Suyash0207

Suyash’s AI Stack

anthropic apis
anthropic apis
claude code
claude code
cursor
cursor
langchain
langchain
langgraph
langgraph
langsmith
langsmith

Education

Birla Institute of Technology, Mesra

Bachelor of Technology - BTech, Computer Science

Birla Institute of Technology, Mesra

20204 yrs2024

Join Suyash and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Suyash Sanjay on Tal

Headline

Founding Engineer @ Anko AI | Building LLM Agents, MCP Integrations & Multi-Agent Systems that Scale

About

I build the agent infrastructure behind production LLM systems, multi-agent orchestration, memory pipelines, and MCP integrations that let AI agents actually learn and act.

Location

Bengaluru

AI stack

anthropic apis

claude code

cursor

langchain

langgraph

langsmith

Work experience

Founding Engineer at Anko AI, Jul 2024 - Present · 2 yrs 2 mos, Bengaluru, - Engineered a human-in-the-loop memory pipeline: an LLM agent extracts structured memory candidates from historical chat sessions, validates them against a live entity registry, and promotes approved facts into future analyst prompts — enabling continuous agent learning. - Designed and built end-to-end MCP integration enabling secure external client access to a production chat platform, including session persistence, token-based authentication, and cross-service request routing. - Built a production LangGraph functional FSM for multi-turn LLM agent execution with SSE streaming, tool validation, reasoning/response separation, and runaway-loop safeguards. - Designed a hierarchical multi-agent system where sub-agents are invoked as structured tools from a parent Analyst agent, enabling compositional analytics reasoning. - Implemented LLM-driven code execution tools allowing agents to dynamically write and run Python/pandas scripts and generate Word reports within the chat loop. - Built an internal prompt engineering platform (Eva) with versioned Jinja2 templates, auto-discovered MKG variable injection, and a dynamic skills system that injects new prompts and tools into live agent executions at runtime. - Migrated the ThoughtSpot integration to the official REST API SDK with stateless bearer token auth and automatic 401 retry, providing a reliable data query layer for LLM agents at runtime. - Designed and architected the frontend with a service layer and modular design, reducing build times by 50% and producing a codebase structure that LLMs can navigate and extend reliably.

SWE Intern at Anko AI, Jan 2024 - Jul 2024 · 7 mos, Bengaluru, Karnataka, India, • Spearheaded the development of a new frontend service for the MVP, which resulted in the company's first paid customer. • Designed and built an admin panel for metric setup and visualization using Vis.js, reducing manual setup time by 30%. • Upgraded the frontend service from JavaScript to TypeScript React, improving type safety, maintainability, and scalability. • Implemented a Snowflake integration panel to create and update data models, cutting data integration time by 70%.

Software Engineer Intern at Rachnaye, Jun 2023 - Jul 2023 · 2 mos, Bengaluru, Karnataka, India

Education

Bachelor of Technology - BTech, Computer Science, Birla Institute of Technology, Mesra

Public links

leetcode: https://leetcode.com/Suyash0207