Chapter 00 · Origin
It began with signals, not software.
An electronics & communication engineer by training, Shoeb's first job was reading the invisible — drive-test data of radio towers at Ada Cell Technologies. Every spectrum plot was a pattern waiting for a model. That curiosity never left; it only got bigger inputs.
0Years of experienceFrom RF-image analysis at telecom giants to agentic systems today.
0Brands servedAcross telecom, healthcare, insurance, energy, retail and automotive.
0Projects deliveredComputer vision, generative AI and agentic systems.
0Platforms ownedEnd-to-end AI platforms architected and owned.
Alumni & academic reach
IITIIT MadrasMS, Artificial Intelligence
BITSBITS PilaniM.Tech, Artificial Intelligence
Guest speaker at
IIM AhmedabadIIT DelhiISROPES UniversityPyDataSITAM
Chapter 05 · The Agentic Turn
Then models learned to talk. Shoeb taught them to collaborate.
From 2023, the same architect-mind moved from single models to societies of agents — orchestrators, routers, graph-driven workflows, agents that discover each other through cards. Enterprise-grade, on Azure, in production-shaped POCs for global brands.
05Lead Architect
ProcurementOne — a platform where workflows are data
The flagship. An enterprise agentic procurement platform unifying buying, should-cost, spend, invoice, PO, contract and ERP.
▸ DAG engine in the database
Agents are nodes, edges are labelled; topological sort yields parallel layers over Azure Service Bus. Change a workflow in the Pipeline Builder — no code release.
▸ A2A microservices · 20+ agents
JSON-RPC 2.0 over HTTP, discoverable through agent cards. Onboarding an agent is configuration, not an orchestrator rewrite.
▸ LLM routing gateway + hybrid retrieval
Routes the first request against live skill cards; pgvector + full-text via Reciprocal Rank Fusion; ERP ingestion in LangGraph.
LangGraphAzure OpenAIService BuspgvectorTerraformA2A
06The enterprise roster
Brands that trusted the graph
▸ Apple Ads — Email Intelligence
Multi-agent platform turning sales email into Hot/Warm/Cold leads, SLA breach alerts and team insights — with auditable structured LLM outputs.
▸ BMW Group — Change Request Management
Orchestrator, SAP, RAG, OCR and Ingestion agents automating change-request workflows end to end.
▸ Al-Futtaim — Procurement Transformation
Guided Buying, Recommendation and Negotiation agents with measurable cycle-time reduction.
▸ PepsiCo · BHP · MetLife · GEM
Five-agent NL-to-SQL analytics with graph ontology; persona-aware knowledge sharing; GenAI retirement messaging; BERT-based content moderation.
Chapter 07 · Knowledge Graph
Everything above, as one connected mind.
Careers are graphs, not timelines. Drag nodes, hover to light up connections, click any node to explore how companies, domains, projects and technologies relate.
PICK A TAB · CLICK A NODE TO EXPAND ITS CONNECTIONS · DRAG TO MOVE
