
CLOUD ARCHITECTURE
MY PASSION FOR CLOUD
Having a wild imagination and traits of perfectionism, I had struggled with finding correct resources and time to bring my imaginations to life. Comes the cloud platforms, Google Cloud Platform and Amazon Web Services Platform, getting access to a wide variety of tools abled me to develop tools and with ease.
Cloud Platform is the world at your fingertips
​
​
Cloud Architecture is the blueprint of any tool developed in cloud. It includes several key components and multiple sub tools offered by cloud platforms for specific uses. A properly built architecture is valuable for productivity gains, less costly to maintain and less time consuming. Interact with the following App to understand critical components and my favorite tools being used for my projects.
Click on the main circles.
Projects Portfolio
My Projects Portfolio include the most recent projects that I have worked with and been a part of. Feel free to interact and get to know more about the projects. The projects include the project architecture, tools being used, projects details, skills and certifications acquired, my main responsibilities. Some details for some projects are classified under NDAs hence can not be revealed.
Cloud Projects​​
Automated Org Chart Builder for AMARIS
Purpose of the project was to Enable AMARIS consultant managers to ingest static team-structure images for third-party clients, enrich personnel records, and maintain a live, exportable organization chart. As only pictures of client team structures were present and lacked an automated way to capture roles, contact details, LinkedIn profiles, and bios.
High-Level Outcome: Reduced manual org-chart creation time from 2 months to under 7 days and improved data-update cadence to near real time.
Business Objectives
Accelerate go-to-market for org-chart deliverables
Ensure data freshness with continuous updates
Enrich basic hierarchies with complete contact metadata
Achieve 95%+ data extraction accuracy
Automate end-to-end processing within a 7-day sprint
Securely manage sensitive personal data
Solution Design
Data Flow & Control Flow:
-
User selects company & uploads org-chart image via UI
-
UI stores image in Google Cloud Storage
-
“DocProc” fetches image, calls the custom Document AI processor, extracts hierarchy (Manager, Department, Team Members)
-
Processor writes raw hierarchy into BigQuery
-
“EnrichProc” retrieves new rows, queries Lusha API for emails, phones, LinkedIn, bios, and updates BigQuery
-
Looker Studio renders the hierarchical chart; users export to PDF/Excel
Core Components:
-
UI front-end for image upload (React)
-
Google Cloud Storage bucket (raw images)
-
Cloud Run service “DocProc” invoking custom Document AI model
-
BigQuery dataset “org_charts”
-
Cloud Run service “EnrichProc” calling Lusha API
-
Looker Studio dashboard with live BigQuery connection
Scaling & Compliance:
-
Cloud Run autoscaling (min-instances=1, max-instances=10)
-
Multi-region BigQuery with default replication
-
IAM roles:
-
UI service account with “storage.objectViewer”
-
DocProc service account with “documentai.process” & “bigquery.dataEditor”
-
EnrichProc service account with “bigquery.dataEditor”
-
-
VPC Service Controls around Storage & BigQuery
-
Cloud Armor firewall rules on all ingress
-
Data encrypted at rest and in transit
Results
CV Skills Analyzer
Streamline AMARIS recruitment process through automated CV analysis, skills extraction, and intelligent candidate-job matching to eliminate manual screening bottlenecks.
High-Level Outcome: Reduced candidate screening time from 2 weeks to instant while improving match accuracy by 85% and establishing a scalable recruitment intelligence platform.
Accelerate time-to-hire by eliminating manual CV review processes
Improve candidate-job matching precision through standardized skills analysis
Establish data-driven recruitment insights and reporting capabilities
Achieve 95%+ accuracy in CV data extraction and skills identification
Process 500 24 hours
Implement real-time candidate ranking with weighted suitability scores
Build secure, compliant data handling for sensitive candidate information
Business Objectives
Solution Design
Core Components:
-
React-based UI for consultant managers
-
Cloud Run "CVProcessor" service with custom Document AI model
-
Cloud Run "SkillsExtractor" integrating Lightcast API
-
Cloud Run "JobAnalyzer" for job requirement processing
-
Custom Vertex AI Gemini model for candidate matching
-
BigQuery data warehouse with structured candidate/job datasets
-
Looker Studio dashboard for recruitment analytics
Data Flow & Control Flow:
-
Manager enters Job ID into UI
-
"CVProcessor" calls Mantu ATS API to retrieve all candidate CVs for position
-
Document AI OCR extracts structured data (name, email, mobile, experience, languages)
-
"SkillsExtractor" sends CV text to Lightcast API for skills identification and normalization
-
Skills data stored in BigQuery with nested JSON structure
-
"JobAnalyzer" fetches job description from Mantu database via API
-
Job requirements processed through Lightcast for required skills taxonomy
-
Custom Gemini model analyzes candidate-job fit using proprietary framework
-
Weighted suitability rankings generated based on must-have/hard/soft skills matching
-
UI displays ranked candidates with detailed skills breakdown and scores
Scaling & Compliance:
-
Cloud Run auto-scaling (min-instances=2, max-instances=50 per service)
-
BigQuery multi-region replication with 99.9% SLA
-
Circuit breaker patterns for external API dependencies
-
IAM roles:
-
UI service account: mantu.ats.reader, bigquery.dataViewer
-
CVProcessor: documentai.process, bigquery.dataEditor
-
SkillsExtractor: bigquery.dataEditor, external API access
-
JobAnalyzer: bigquery.dataEditor, Mantu API access
-
Gemini service: aiplatform.user, bigquery.dataViewer
-
-
VPC Service Controls around BigQuery and sensitive data stores
-
Cloud Armor security policies for API protection
-
Data encryption at rest and in transit with customer-managed keys
Results
Automated Social Media Contest Winner Selection System
Streamline Freedom Confectionery's social media contest management by automating the tracking of Instagram follows, hashtag usage, and Google reviews to eliminate manual winner selection processes
High-Level Outcome: Transformed a 4-day manual process into a 1-day automated system, reducing marketing team workload by 80% while maintaining compliance with privacy regulations and NDAs
Business Objectives
Accelerate contest completion cycles to maintain social media engagement momentum
Reduce manual verification workload for marketing team members
Ensure fair and transparent winner selection through automated processes
Scale contest frequency without proportional increase in human resources
Achieve 75%+ accuracy in cross-platform participant identification
Process contest entries within 24 hours of challenge completion
Maintain data privacy compliance while handling social media information
Build scalable architecture supporting multiple concurrent contests
Solution Design
Core Components:
-
Web-based UI for contest management and winner announcement
-
BigQuery data warehouse with separated contestant action databases
-
Custom Vertex AI Gemini model for cross-platform identity matching
-
Cloud Run "InstagramProcessor" service utilizing Instagram REST API
-
Cloud Run "ReviewsCollector" integrating Google Reviews API
-
Cloud Run "WinnerSelector" for automated contest result generation
-
Looker Studio dashboard for analytics
Data Flow & Control Flow:
-
Due to NDA, the Data Flow and Control Flow can not be revealed
Scaling & Compliance:
-
Cloud Run auto-scaling (min-instances=1, max-instances=20 per service)
-
BigQuery partitioned tables by contest date for performance optimization
-
Circuit breaker patterns for Instagram and Google API dependencies
-
IAM roles with principle:​
-
VPC Service Controls around BigQuery and sensitive data stores
-
Cloud Armor security policies for API protection
-
Data retention policies compliant with privacy regulations
Results
MY EXPERTISE TOOLS
Following are the tools and skill areas I am specialized in. However, these don't limit me.
Data Analysis
Data Visualization
Data Analytics
R
Data Cleaning
SQL
API Dev.
Web Services
Programming
Containerization
Cloud Computing
Google BigQuery
Kubernetes
GCP
Terraform
AWS
AZURE
Docker
Cloud Architecture
Python
Java Script
JSON
DevOps
PowerShell
Cloud Automation
IAM
Vertex AI
Document AI
Gemini AI
VPC
Amazon S3
Virtual Machines
Azure Storage
CI/CD Pipeline
GKE
Jenkins
ELK Stack
Google Cloud Logging
RBAC
ABAC
MFA
SSO
GDPR
CCPA
DNS Configurations
Google Cloud SQL
Firestone
YAML
AGILE
Project Management
SCRUM
JIRA
MIRO
Trello
Machine Learning
MLOps
Edge Computing
Cost Optimization
MY CERTIFICATIONS
Click the floating icons to get to know more in details of the certifications
WANT TO GET TO KNOW MORE?
Why not contact me?
Or you can ask the AI(Digital Version of Me) associated trained with real life data and keep getting updated.
The complete website is automatically getting updated from every little experience the physical me experience day to day life. The platform is getting updated every day. Keep in touch and