Loading
AI Engineer
Adil Qureshi
Generative AI Engineer
I build production-grade Generative AI and Agentic AI systems powered by Gemini, Vertex AI, FastAPI, and Google Cloud Platform.
Specialized in multi-agent orchestration, RAG, document intelligence, event-driven cloud pipelines, structured AI output, and explainable AI applications.
GeminiVertex AIFastAPIPythonGoogle CloudBigQueryCloud RunPub/SubDocument AIAgentic AI
Featured Work
Enterprise Agentic AI, shipped on Google Cloud
A selection of production systems built with Gemini, Vertex AI, and event-driven cloud architecture.
Enterprise Agentic AINorthwell HealthGemini and Vertex AI8+ GCP ServicesExplainable Recommendations
PRReviewAgenticAI
Enterprise Agentic AI for procurement requisition review
GeminiVertex AIFastAPIPythonCloud RunPub/Sub
An event-driven, multi-agent system that ingests procurement requisition data and documents, reasons over them with Gemini and Vertex AI, and produces explainable, confidence-scored recommendations surfaced in a Recommendation Hub dashboard — built for Northwell Health on Google Cloud.
Agent Workflow Orchestrator
Composable multi-agent orchestration layer
Vertex AIGeminiFastAPIPub/SubCloud Run
A reusable orchestration layer that classifies, validates, and routes work between specialized agents, coordinating Gemini reasoning calls with deterministic business rules and structured output contracts.
Document Intelligence Pipeline
Automated document understanding at scale
Document AICloud StorageBigQueryPython
A Document AI powered pipeline that extracts, classifies, and structures unstructured procurement documents before handing them to downstream reasoning agents.
Cloud-Native AI Architecture
Event-driven infrastructure for agentic systems
TerraformCloud RunPub/SubCloud Monitoring
Infrastructure-as-code blueprint for deploying multi-agent AI systems on Google Cloud — Pub/Sub triggers, Cloud Run services, Firestore state, and full observability via Cloud Logging and Monitoring.
How It Works
Cooperating agents, one explainable recommendation
Hover or focus any node to see its responsibility, inputs, outputs, and the technology behind it.
- 1. Requisition Data
- Source procurement requisition records enter the system.
- In: Raw requisition + attached documents → Out: New requisition event
- Cloud Storage
- 2. Pub/Sub Trigger
- Publishes an event the moment a new requisition arrives.
- In: New requisition event → Out: Pub/Sub message
- Pub/Sub
- 3. Cloud Run
- Scales the processing service up on demand to handle the event.
- In: Pub/Sub message → Out: HTTP request to API layer
- Cloud Run
- 4. FastAPI
- Validates the request and routes it into the agent system.
- In: HTTP request → Out: Orchestration call
- FastAPI
- 5. Agent Orchestrator
- Coordinates the specialized agents and tracks pipeline state.
- In: Requisition payload → Out: Task assignments
- Python
- 6. Classification Agent
- Determines requisition type, urgency, and routing category.
- In: Requisition payload → Out: Category + priority label
- Gemini
- 7. Validation Agent
- Checks completeness, policy compliance, and budget constraints.
- In: Category + priority label → Out: Validation result
- Vertex AI
- 8. Document Agent
- Extracts and structures data from attached source documents.
- In: Attached documents → Out: Structured document data
- Document AI
- 9. Recommendation Agent
- Synthesizes agent outputs into an actionable recommendation.
- In: Validation result + document data → Out: Draft recommendation
- Gemini + Vertex AI
- 10. Gemini and Vertex AI
- Performs the core reasoning and generates explainable rationale.
- In: Draft recommendation → Out: Reasoned recommendation + rationale
- Gemini, Vertex AI
- 11. Confidence Score
- Scores the recommendation's reliability for human reviewers.
- In: Reasoned recommendation → Out: Confidence-scored recommendation
- Vertex AI
- 12. Recommendation Hub
- Surfaces the final recommendation to reviewers in a dashboard.
- In: Confidence-scored recommendation → Out: Reviewed decision
- Firestore + Firebase Auth
Cloud Architecture
Event-driven infrastructure on Google Cloud
PRReviewAgenticAI runs across 8+ managed Google Cloud services, provisioned entirely with Terraform.
Ingestion
Cloud Storage
Pub/Sub
Compute & AI
Cloud Run
Document AI
Vertex AI
Gemini
Data & Access
Firestore
BigQuery
Firebase Authentication
Observability
Cloud Logging
Cloud Monitoring
Every environment above is provisioned and managed with Terraform.
Engineering Process
From problem to production
The workflow behind every AI system Adil ships.
- 1
Problem Understanding
Working with stakeholders to define the real business problem, constraints, and success criteria before writing any code.
- 2
Data and Document Ingestion
Bringing structured data and unstructured documents into the pipeline via Cloud Storage and Document AI.
- 3
Prompt Engineering
Designing prompts and structured output schemas that make Gemini's reasoning reliable and explainable.
- 4
Multi-Agent Orchestration
Composing specialized agents — classification, validation, document, recommendation — into a coordinated pipeline.
- 5
Gemini Reasoning
Letting Gemini and Vertex AI perform the core reasoning, scoring, and rationale generation.
- 6
FastAPI Integration
Exposing agent capabilities through typed, testable FastAPI services.
- 7
Google Cloud Deployment
Deploying event-driven services on Cloud Run, wired together with Pub/Sub and managed with Terraform.
- 8
Monitoring and Improvement
Observing pipeline health with Cloud Logging and Monitoring, then iterating on prompts and agent logic.
Skill Constellation
An interactive map of the stack
Click a category to expand the technologies underneath it.
Adil Qureshi
AI Engineer
Credentials
Certifications
Formal recognition backing the hands-on Agentic AI and cloud engineering work above.
Agentic AI
Data Science
IBM
Envisioning Your Solution
IBM
Try It
AI Terminal
A sandboxed terminal — click a command or type your own. Nothing here touches a real shell.
adil@ai-engineer:~
Welcome. Type `help` or click a command below.
Gallery
Architecture, dashboards, and agent workflows
A closer look at the systems behind PRReviewAgenticAI.
Get In Touch
Let's build something explainable
Open to enterprise Agentic AI, Generative AI, and Google Cloud engineering opportunities.