Sabya Sachi profile photo 1 on Tal

Sabya Sachi

Engineering Manager @Omnicom | Agentic AI | RAG | LLM | LangGraph | Python | Simplifying learning & Sharing Insights on AI | Designing & Thinking in Systems | Web Development

Bengaluru

10+ years of work experience

Omnicom Media
Dec 20232 yrs 9 mosPresent

India

Architected an event-driven, AI-agent-powered dashboard automation platform on a unified semantic model, scaled to ~70 clients — live reporting with instruction-based Copilot experiences in Power BI and Tableau, grounded in governed metric definitions. Designed and shipped custom dashboard data agents exposed via MCP, making governed analytics data queryable by AI assistants across the application; built Python FastAPI agent and integration endpoints backing these experiences. Built an embedded conversational analytics experience — a custom in-dashboard AI chat panel with streaming responses, calling data-agent endpoints directly, enriched with real-time report context (active page, filters, and selections captured via the Power BI client SDK and streamed to a FastAPI backend) so answers are grounded in exactly what the user is viewing. Established an AI-augmented frontend SDLC: governed prompt libraries, agent project configurations (CLAUDE.md), and MCP servers wired into engineers' environments so AI-generated components conform to the design system, accessibility standards, and coding conventions by default. Pioneered AI-driven UI development across 50 engineers: MCP-driven component generation from a governed design system, Figma-to-code workflows enforced through design tokens and UI contracts, and prompt-engineered scaffolding for new modules — reducing build time for standard dashboard and platform screens by ~40%. Embedded LLM-powered code review and quality automation (Claude Code, GitHub Copilot) into the engineering workflow — AI-assisted PR review, standards enforcement, and test generation — raising code quality and velocity across all product tracks. Scaled the web engineering team from 5 to 50 engineers, structuring them across product tracks (dashboards, reporting, analytics integrations, platform modules, design system) with consistent standards, DevEx, and delivery workflows.
Annalect India
Jan 202311 mosNov 2023

Bengaluru, Karnataka, India

Annalect India
Apr 20211 yr 10 mosJan 2023

Bengaluru, Karnataka, India

emids
Jul 202010 mosApr 2021

Greater Bengaluru Area

emids
Jul 20182 yrs 1 moJul 2020

Greater Bengaluru Area

Full Stack Developer
Dotdash Meredith
Jun 20171 yr 2 mosJul 2018

Bengaluru, Karnataka, India

Designed end-to-end financial data processing pipelines ingesting mainframe data, applying validation, and persisting to SQL Server; automated scheduled processes via Windows Services. Built invoice-generation modules with analytical transformations over subscriber data and dynamic Crystal Reports; owned requirements, design, and delivery across full SDLC. Received an individual award for the daily invoicing module.
Empower
Jun 20161 yrMay 2017

Enhanced the retirement Withdrawal Module and built the AngularJS-based Advisory Module for American Funds, including async promise-based workflows and web-service integration between UI and business layers. Modeled retirement-plan user types in Oracle and delivered POCs validating architecture for upcoming modules; received an individual award for the withdrawal UI module.

Sabya’s AI Stack

claude code
claude code
claude
claude
cursor
cursor
github copilot
github copilot
langgraph
langgraph
langsmith
langsmith

Education

IET Alwar

Bachelor of Technology, Computer Engineering

IET Alwar

20074 yrs2011
Rajasthan Technical University, Kota

B.Tech, Computer Science & Engineering

Rajasthan Technical University, Kota

20074 yrs2011

Join Sabya and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Sabya Sachi on Tal

Headline

Engineering Manager @Omnicom | Agentic AI | RAG | LLM | LangGraph | Python | Simplifying learning & Sharing Insights on AI | Designing & Thinking in Systems | Web Development

About

› 14+ years of experience building enterprise platforms, analytics products, and AI-native applications across finance, healthcare, media, and marketing domains. › Architected an event-driven, AI-agent-powered analytics platform on a unified semantic model serving 50+ clients, with custom MCP data agents bringing governed analytics into AI assistants across the organization. › Pioneered AI-driven UI development — MCP-based component generation, Figma-to-code workflows governed by design tokens, and LLM-powered code review (Claude Code, GitHub Copilot). › Lead an embedded-BI platform (Power BI, Tableau, Datorama) serving ~20K users (8K DAU) for Volkswagen, Nissan, Meta, Amazon, and PepsiCo. › Hands-on engineering leader with expertise spanning Agentic AI, Power BI, Tableau, Fast API, AWS, distributed systems, frontend architecture, and platform engineering.

Location

Bengaluru

AI stack

claude code

claude

cursor

github copilot

langgraph

langsmith

Work experience

Engineering Manager at Omnicom Media, Dec 2023 - Present · 2 yrs 9 mos, India, Architected an event-driven, AI-agent-powered dashboard automation platform on a unified semantic model, scaled to ~70 clients — live reporting with instruction-based Copilot experiences in Power BI and Tableau, grounded in governed metric definitions. Designed and shipped custom dashboard data agents exposed via MCP, making governed analytics data queryable by AI assistants across the application; built Python FastAPI agent and integration endpoints backing these experiences. Built an embedded conversational analytics experience — a custom in-dashboard AI chat panel with streaming responses, calling data-agent endpoints directly, enriched with real-time report context (active page, filters, and selections captured via the Power BI client SDK and streamed to a FastAPI backend) so answers are grounded in exactly what the user is viewing. Established an AI-augmented frontend SDLC: governed prompt libraries, agent project configurations (CLAUDE.md), and MCP servers wired into engineers' environments so AI-generated components conform to the design system, accessibility standards, and coding conventions by default. Pioneered AI-driven UI development across 50 engineers: MCP-driven component generation from a governed design system, Figma-to-code workflows enforced through design tokens and UI contracts, and prompt-engineered scaffolding for new modules — reducing build time for standard dashboard and platform screens by ~40%. Embedded LLM-powered code review and quality automation (Claude Code, GitHub Copilot) into the engineering workflow — AI-assisted PR review, standards enforcement, and test generation — raising code quality and velocity across all product tracks. Scaled the web engineering team from 5 to 50 engineers, structuring them across product tracks (dashboards, reporting, analytics integrations, platform modules, design system) with consistent standards, DevEx, and delivery workflows.

Associate Manager at Annalect India, Jan 2023 - Nov 2023 · 11 mos, Bengaluru, Karnataka, India

Lead Product Developer at Annalect India, Apr 2021 - Jan 2023 · 1 yr 10 mos, Bengaluru, Karnataka, India

Technical Team Lead at emids, Jul 2020 - Apr 2021 · 10 mos, Greater Bengaluru Area

Full-stack Developer at emids, Jul 2018 - Jul 2020 · 2 yrs 1 mo, Greater Bengaluru Area, Full Stack Developer

Technical Lead at Dotdash Meredith, Jun 2017 - Jul 2018 · 1 yr 2 mos, Bengaluru, Karnataka, India, Designed end-to-end financial data processing pipelines ingesting mainframe data, applying validation, and persisting to SQL Server; automated scheduled processes via Windows Services. Built invoice-generation modules with analytical transformations over subscriber data and dynamic Crystal Reports; owned requirements, design, and delivery across full SDLC. Received an individual award for the daily invoicing module.

Senior Software Engineer at Empower, Jun 2016 - May 2017 · 1 yr, Enhanced the retirement Withdrawal Module and built the AngularJS-based Advisory Module for American Funds, including async promise-based workflows and web-service integration between UI and business layers. Modeled retirement-plan user types in Oracle and delivered POCs validating architecture for upcoming modules; received an individual award for the withdrawal UI module.

Education

Bachelor of Technology, Computer Engineering, IET Alwar

B.Tech, Computer Science & Engineering, Rajasthan Technical University, Kota