Vishvas Janardhan

Most RAG systems are demos.Mine runs on a 7.5 TB estate I administer myself.

I lead the development team at an engineering consultancy in Oman, where I build LLM retrieval over project data — and own the database cluster underneath it. Eleven years of production .NET, React and MSSQL, the most recent of them spent putting applied AI in front of people who use it daily.

AI & Platform Engineering
Muscat · GST (UTC+4)
contact@vishj.comlinkedin.com/in/vishjgithub.com/vish-jResume on request →
11 yrs
building enterprise systems
7.5 TB
SQL Server estate administered
550+
daily users on platforms I lead
<30 s
document retrieval, from ~15 min
01

Selected work

01

Retrieval over engineering project data

Natural-language search across 125,000+ documents and a decade of project history, built on a .NET retrieval pipeline with Azure OpenAI and vector search.

Azure OpenAIAzure AI Search.NET 8MSSQL
~15 min → under 30 s
Read →
02

A 7.5 TB SQL Server estate

Two-node SQL Server 2022 Enterprise active-active cluster carrying 80+ production databases, migrated forward from 2012 across four platform generations without losing a maintenance window.

SQL Server 2022Always On AGFCITDEPowerShell
80+ databases · RPO under 15 min
Read →
03

Document control platform

The system the retrieval layer reads from: a multi-stage review and approval platform serving 1,500+ live projects, architected and delivered with a team of three.

ASP.NET CoreReactMSSQLHangfire
550+ users · 1,500+ projects
Read →
02

Experience

Feb 2021 — Present

Senior Software Developer & Team Lead

Muscat Engineering Consultancy · Engineering consultancy, Oil & Gas and infrastructure
  • Lead a team of three developers across all in-house enterprise platforms, owning solution architecture, code review standards and release planning.
  • Architected the document control platform in ASP.NET Core, React and MSSQL — now 550+ active users, 1,500+ live projects, 125,000+ documents and 350,000+ versions.
  • Built the LLM retrieval layer over project data, and administer the two-node SQL Server 2022 cluster it runs on: 80+ databases, 7.5 TB, encrypted at rest.
Mar 2015 — Feb 2021

Software Developer

Bilfinger Tebodin · International engineering consultancy, Oman operations
  • Bootstrapped the company's business platform from nothing in 2015 and grew it into 20+ modules used by up to 900 staff.
  • Split the monolith into a multi-project solution, moved the data layer to EF Code-First, and introduced the first unit-test suite.
  • Delivered the e-document management system (450 users) and the purchasing system (900 users) end to end, from requirements through support.
Sep 2014 — Feb 2015

ICT Intern

Tebodin & Partner · Muscat, Oman
  • Automated company-wide application deployment and configured an ELK stack for centralized log analysis and audit reporting.
2011 — 2015
BEng, Data Communication and System Administration
Coventry University · Graduated with Honours in Data Communication
2009 — 2011
Diploma, Multimedia Technology
Coventry University
03

Tools

AI engineering

Azure OpenAI Service · RAG pipelines · Vector embeddings · Azure AI Search · LLM integration · Workflow automation

Backend

.NET / C# · ASP.NET Core · Entity Framework Core · REST API design · Hangfire

Frontend

React · TypeScript · JavaScript

Data & infrastructure

SQL Server 2012–2022 · T-SQL · Query tuning & indexing · Always On availability groups · Failover cluster instances · TDE · Backup & point-in-time recovery · Capacity planning

Platform

Azure App Service · Azure SQL · Azure DevOps CI/CD · PowerShell · Git · ELK · Power BI · Dynamics 365

Leadership

Team leadership · Solution architecture · Code review standards · Agile delivery · Mentoring · Technical hiring

04

Get in touch

Happy to talk about retrieval systems, SQL Server at scale, or what it takes to put an LLM in front of data people actually depend on.

contact@vishj.com
LinkedIn · GitHub · Resume on request