Named outcomes, honest labels.
Enterprise engagements are labeled “via” the contracting partner. Founder experience that predates QueUp is marked as such — never claimed as QueUp client work.
RAG agents for automated task creation — Microsoft internal cloud platform (via Akvelon)
Engineering teams were manually reviewing completed work to plan follow-up tasks and assignments — slow, inconsistent, and unscalable across a global platform.
We researched and implemented RAG-based agents on the OpenAI platform, fed by ETL/background jobs that analyzed completed work and generated structured tasks and team assignments automatically. In a second phase we built the data backbone: integrations aggregating operational data over Kusto (KQL), SQL, GraphQL, and HTTP, surfaced through dashboards used by engineering and product leadership.
1TB ETL migration: data latency cut from ~48 hours to near real-time
A US healthcare distribution company ran reporting on a legacy ETL pipeline that took roughly two days to surface operational data — decisions were always made on stale numbers.
We redesigned the pipeline end to end: incremental ingestion, restructured staging, and cloud-native processing on Azure, migrating ~1TB of production data without interrupting reporting for the business.
Vendor-to-internal platform handover without losing velocity
Tenderr needed to take a vendor-built platform in-house — with full knowledge transfer, no regression in delivery pace, and a codebase its own team could carry forward.
We stabilized the system, documented the architecture, restructured the delivery pipeline, and ran a staged handover so the internal team took ownership incrementally rather than in one risky cut-over.
AI document workflow for a German research institute
Researchers were manually extracting structured information from large volumes of documents — repetitive work that consumed skilled staff time.
We built a document-automation pipeline combining Textract-based extraction with GPT-driven structuring and validation, integrated into the institute's existing workflow.
BaitulMal — donation-tracking fintech, designed, built, and operated by QueUp
Non-profits struggle to give donors transparent, real-time visibility into where money goes — most tooling is spreadsheets or opaque accounting exports.
We designed, built, and operate the full platform: .NET/Azure backend, React web app, React Native mobile app, and a white-label donation website offered as an add-on. Same team, same standards our clients get.
Java Fashion — online menswear store, set up and launched end to end
A menswear retailer with no technical staff wanted to sell online across Serbia — prices in RSD, cash on delivery, and a store they could run day to day without a developer.
We did all of it: Shopify account and domain, a customized theme, catalog and categories, a Serbian-language storefront, checkout and notifications, shipping rules with free delivery over 10,000 RSD, cash-on-delivery payments, promotions, and store policies — then kept maintaining it after launch.
In their words.
“BaitulMal has transformed how we manage donor funds and track expenses across multiple projects. The reporting features have saved us countless hours when preparing for audits.”
