Retail · Full System
Fashion Boutique · Riga · Physical store + WooCommerce
Inventory sync and management portal for a Riga fashion boutique
A women's clothing boutique managing stock entirely in Excel — staff updated it manually after each sale, independently, which meant it was constantly wrong. The result: online orders for items that weren't in stock, angry customers, and hours every week reconciling discrepancies. They'd looked at off-the-shelf solutions but none handled their specific combination well, and the monthly SaaS fees added up.
What we built
A Full System: a purpose-built SQL Server database, web-based stock management portal, and integration with their existing WooCommerce store.
- SQL Server database: products, variants (size/colour), stock levels, supplier records, sales history, customer records
- Web-based management portal (browser-based, works on tablets at the counter) for stock intake, counts, and adjustments
- WooCommerce sync: stock updates automatically within 60 seconds of an online sale
- Low-stock alerts emailed when any variant drops below a configurable threshold
- Supplier purchase orders: create POs in the portal, receive stock against them, automatic stock uplift on receipt
- Daily sales report emailed automatically each morning: top sellers, margin by category, stock health
0Oversells/month
3 hrsWeekly admin saved
23%Margin improvement
6 wksDelivery time
C# .NET
SQL Server
WooCommerce API
ASP.NET Core
Dapper
Full System
Healthcare · Custom Software
Dental Clinic · 3 Locations · Riga region
Patient management system for a 3-location dental clinic
Patient records on paper at each clinic. Appointments in a shared Outlook calendar that three receptionists edited simultaneously — double-bookings were a recurring problem. Billing in Excel, invoiced manually, reconciled at month-end. They'd been quoted €40,000+ for enterprise dental software, plus a per-seat monthly licence that would grow every year. They needed something that fit their actual workflow, not a platform designed for 50-location chains.
What we built
A web-based patient management system shared across all three locations in real time, with role-based access for reception, dentists, and management.
- Central SQL Server database: patients, appointments, treatments, invoices, and staff across all three locations
- Patient records module: medical history, contact details, treatment notes, photo attachments, GDPR-compliant
- Appointment scheduler: calendar view per dentist per location, conflict detection, SMS/email reminders
- Billing module: treatment-linked invoices, payment recording, outstanding balance tracking, monthly revenue reports
- Role-based access: receptionists see scheduling and billing; dentists see patient records; management sees everything
- Runs in a browser — no software installation, works on existing computers at each clinic
100%Digital records
ZeroDouble bookings
40%Less admin time
€40kEnterprise cost avoided
C# .NET
SQL Server
ASP.NET Core
Multi-location
GDPR
Custom Software
Distribution · Custom Software
Import & Distribution Company · Baltic region
Supplier pricing automation for a Baltic import company
A company importing goods from 14 European suppliers was drowning in price lists — each sent differently: Excel, PDF, CSV, embedded in emails. A staff member spent 3+ days every month downloading, reformatting, comparing, and manually entering prices. Mistakes were common. Margin decisions were made on data that was weeks old.
What we built
An automated supplier price management platform that ingests price lists in any format, normalises them, and delivers clean margin data to the sales team.
- SQL Server database: suppliers, product catalogue, price history, exchange rates, margin rules
- Price import engine (C#): handles Excel (.xlsx/.xls), CSV, and structured PDFs — each supplier's format configured once, then automated
- Normalisation layer: maps each supplier's column names to a standard product catalogue across different SKUs and packaging sizes
- Price comparison dashboard: current price vs. previous price per supplier — highlights changes above configurable thresholds
- Margin calculator: applies exchange rates, import costs, and margin targets to produce suggested retail prices
- Historical price charts: 24-month trend per product per supplier
3d→4hMonthly pricing work
Real-timePrice visibility
14Suppliers automated
24 moPrice history
C# .NET
SQL Server
Excel/CSV/PDF parsing
Data pipeline
Dapper
Banking · Regulatory Compliance
Financial Institution · C#/.NET & SQL Server
Building an automated SEC securities lending reporting system from scratch
A major financial institution needed to comply with new FINRA securities lending reporting requirements. An existing SFTR system was in place, but nothing for SEC. The deadline was tight, the business logic was complex (20+ reporting scenarios), and there was zero room for error — T+1 deadlines with potential fines for non-compliance.
What we built
The complete system: database schema, C# processing engine, JSON file generation, SFTP transmission, and an 881-test automated test suite.
- Normalised database schema with dedicated SEC tables linked to existing SFTR data to avoid duplication
- Event-driven C# processing engine handling 20+ distinct reporting scenarios
- MD5 hash comparison for detecting actual data changes vs. redundant updates
- Intelligent report type determination engine mapping SFTR actions to FINRA report types
- BZip2-compressed JSON file generation with SFTP auto-transmission
- Batch processing schedule aligned to trading hours (13:30, 16:30, 23:00)
- 881 automated xUnit tests with parallel execution and CI/CD in Azure DevOps
100%On-time compliance
~40 hrsManual work/week cut
<5 minBatch processing
ZeroViolations
C# .NET 6.0
SQL Server 2019
Dapper ORM
xUnit
Azure DevOps
Hangfire
SFTP
BZip2
Banking · SQL Optimisation
Financial Institution · Regulatory Reporting · SQL Server
Reducing SFTR report generation from 15 minutes to under 2 minutes
The main aggregation query was performing 47 table scans across tables with millions of rows. The stored procedure had grown organically over 4 years — every new regulation meant another subquery bolted on. Nobody had ever looked at the full picture.
What we did
- Replaced 12 correlated subqueries with indexed views for frequently accessed data combinations
- Implemented incremental processing — only recalculating data that changed since the previous run
- Refactored nested queries into well-indexed temporary tables with filtered indexes
- Moved business logic calculations to persisted computed columns
- Added carefully designed composite indexes aligned with actual query patterns
- Implemented parallel processing for independent data segments
The entire refactoring was done with zero production disruption. Extensive reconciliation testing verified identical results to the original process.
87%Processing time reduced
<2 minReport generation
70%Lower CPU usage
ZeroProduction issues
SQL Server 2019
T-SQL
Indexed Views
Filtered Indexes
Query Tuning