Client Success Stories
Real problems. Real solutions. Real results — from startups to enterprises across the globe.
🏦 Fintech🤖 Generative AI☁️ AWS & DevOps🌍 Global Delivery
Case Studies
How We've Delivered for Our Clients
Two real engagements — different industries, same commitment to quality, speed, and measurable outcomes.
D
Dalstan
Enterprise Fintech Platform — UK & India Operations
↓60%
Infrastructure Cost Reduction
99.9%
System Uptime Achieved
3×
Faster Transaction Processing
Challenge
- Legacy monolithic architecture struggling to handle growing transaction volumes across UK and India markets
- AWS SQS queue failures causing delayed payment processing and customer complaints
- Manual deployments taking 4–6 hours with frequent rollback failures
- AWS costs spiralling with no tagging strategy or cost visibility across teams
- Zero observability — engineers firefighting in production with no monitoring
Solution
- Re-architected core platform on AWS using microservices with ECS and EC2 Auto Scaling
- Fixed SQS queue configuration — corrected IAM roles, dead-letter queues, and visibility timeouts
- Built end-to-end CI/CD pipeline using GitHub Actions — deployments reduced to 12 minutes
- Implemented AWS cost tagging strategy and Reserved Instance planning
- Set up full observability with CloudWatch, Grafana dashboards, and PagerDuty alerting
Impact
- 99.9% uptime sustained across both UK and India production environments
- Transaction processing improved 3× — from 8 seconds to under 2.5 seconds
- Deployment frequency increased from twice a month to daily releases
- AWS bill reduced 60% — saving £40K+ annually
- MTTR reduced from 4 hours to 18 minutes with full observability
G
Genstori
AI-Powered Content & Storytelling Platform — US Market
72hrs
POC to Working Product
↓65%
LLM API Cost Reduction
10×
Content Generation Speed
Challenge
- AI startup with no in-house AI engineering expertise to build their core product
- Direct OpenAI API calls with no caching — costs at $8K/month at low volume
- No RAG pipeline — LLM had no access to user preferences or brand guidelines
- Content generation taking 45–90 seconds per story — too slow for SaaS
- Single EC2 instance with no auto-scaling — at risk during user spikes
Solution
- Built custom RAG pipeline using LangChain and Pinecone connected to user preferences and brand voice
- Implemented semantic caching with Redis — reduced redundant API calls by 65%
- Integrated streaming LLM responses so users see content generating in real-time
- Deployed async job queue with AWS SQS and Lambda for background generation at scale
- Set up auto-scaling ECS cluster with CloudFront CDN for 10,000+ concurrent users
Impact
- Working POC delivered in 72 hours — beta launched within 3 weeks of project kickoff
- LLM costs reduced from $8K/month to $2.8K/month at same usage volume
- Generation time reduced from 45–90 seconds to 4–8 seconds — 10× faster
- Platform handles 10,000+ concurrent users with auto-scaling
- First enterprise deal closed within 60 days citing platform reliability
What Our Clients Say
In Their Own Words
"Vikgol fixed our AWS infrastructure in days — not weeks. The SQS issues that were costing us client relationships were resolved within 48 hours. Their engineers understood our fintech stack immediately and delivered beyond what we expected."
"We went from idea to working AI product in 72 hours. The RAG pipeline Vikgol built is the core of our entire platform. Their team understood generative AI at a depth that's rare — and they moved faster than any agency we'd worked with before."