Open to new projects & collaborations

AI Engineer · Sioux Falls, SD

Alex Jensen

AI Engineer building production systems that actually get used.

Role
AI Engineer
Based in
Sioux Falls, SD
Shipped
10+ systems in production

I build production AI systems — agentic workflows, enterprise RAG, and natural-language data tools — mostly on the Microsoft Agent Framework, with a focus on reliability, privacy, and real adoption. I've shipped 10+ of them into daily use.

View the workGet in touch
Microsoft Agent Framework·Agentic Workflows·Enterprise RAG·NL2SQL·NL2KQL·Real-Time Pipelines·Applied ML·Multi-Agent Systems·Microsoft Agent Framework·Agentic Workflows·Enterprise RAG·NL2SQL·NL2KQL·Real-Time Pipelines·Applied ML·Multi-Agent Systems·
Alex Jensen

About

I build the AI systems people actually keep using.

I build and ship production AI systems — and I've put 10+ of them into daily use across customer care, network operations, field technician workflows, HR, and business intelligence.

Most of it runs on a reusable agent architecture I designed — shared base classes, tooling, and prompt patterns — so new solutions can be stood up quickly. I built it on the Microsoft Agent Framework with Clean Architecture / SOLID across the .NET services, and helped teams adopt and integrate AI into real workflows.

I work across the full stack of an AI system: data and pipelines, model and prompt design, validation and guardrails, applied ML where a model fits better than a prompt, and the product surfaces people actually touch. The hard part is rarely the model — it's concurrency, data privacy, reliability, and adoption.

Selected Work

Case studies in production AI

Most of these live in private company repositories — so here's the architecture and impact behind each. Live demos and walkthroughs available on request.

ANA — Automated Call Notes Assistant

Real-time call summarization for every customer-care call. A service subscribes to a Genesys websocket and logs each call as it ends; a Windows worker service then calls my .NET Web API, pulls the transcript, and runs an LLM summary plus a “Why Customers Call” classification into SQL Server. Built for concurrency — 200 calls per cycle, idempotent, DB-locked against duplicates — with guardrails to keep PII out of every note.

.NET Web APIAgent FrameworkGenesysSQL Server
~30%
less after-call work

Enterprise Knowledge AI Assistant

A company-wide assistant grounded on all internal documentation. An agentic RAG pipeline searches a unified index across ServiceNow, Unily, and a custom Service Desk wiki to answer any employee question — onboarding help, HR-ticket deflection (PTO, benefits), and CSR troubleshooting pulled straight from internal docs.

RAGKnowledge AssistantVector Search
3
doc silos unified

Network Intelligence Assistant

Plain-English access to network and business-intelligence data for leadership. A chatbot with an LLM equipped to query the warehouse through a semantic model — churn and percent-disconnects by area, marketing offers by region / franchise / TOC, top-performing nodes, and performance comparisons — all without anyone writing SQL.

NL2SQLSemantic ModelTool Use
0
SQL required

Modem Anomaly Detector

Pinpoints whether a connectivity issue lives at the modem or the amplifier. An Agent Framework workflow takes a modem by MAC address and analyzes its PNM and Wi-Fi telemetry against the surrounding modems on the same amp. A LightGBM model with isotonic regression outputs a calibrated confidence score, then an LLM turns the metric trends and Z-scores into a plain-language summary for non-technical CSRs.

LightGBMAgent FrameworkPNM Telemetry
ML + LLM
root-cause triage

Log Analytics Agent

Anomaly detection across raw network logs. An agent equipped with a tool to analyze Cisco and network-device logs in Azure Log Analytics — surfacing anomalies and correlated events — and to translate natural language into KQL against the Log Workspace. Deployed in Microsoft Teams and a custom web app for the Log Analytics and Security teams.

NL2KQLMCP ServerAzureTeams
live
log anomaly detection

“Why Customers Call” Dashboard

Rebuilt the CSR account application with AI at the center. The instant an account opens, an LLM summary shows the most likely reasons the customer is calling — reasoned over payment history, past work orders, the last 30 days of call summaries (from ANA), and a week of modem Wi-Fi data — giving reps instant context.

PredictiveLLM ReasoningCSR Tooling
instant
call-reason context

AI Job Description Builder

Modernized 500+ job descriptions, then built the tool to keep doing it. A workflow extracts legacy content, maps it into a new Word template with per-section placeholders, and rewrites section-by-section with human review. The interactive builder (Blazor) walks managers through authoring a new description with AI suggestions, then exports a finished doc and bulk-uploads to ADP.

Agent FrameworkBlazorADPHITL
500+
descriptions modernized

Technician Field Summaries

LLM briefings that prep technicians before they roll a truck. For maintenance, a summary of the amplifier's PNM trends over time; for in-home trouble calls, the customer's issue mined from the transcript captured when the call was scheduled — so techs know what they're walking into.

PNM TrendsTranscript MiningLLM Summaries
pre-truck
field briefings

Capabilities

What's under the hood

AI & LLM

  • Microsoft Agent Framework
  • Agentic & multi-agent workflows
  • Retrieval-Augmented Generation (RAG)
  • NL2SQL / NL2KQL engines
  • Function calling, tool use & MCP servers
  • Prompt engineering, guardrails & PII safety

Engineering & ML

  • C# / .NET · Python · Blazor
  • Clean Architecture & SOLID
  • Concurrent & idempotent pipelines
  • Windows Worker Services · REST APIs
  • LightGBM · isotonic regression
  • Calibrated confidence scoring

Data & Platforms

  • SQL Server · KQL · vector databases
  • Azure Log Analytics · semantic models
  • Genesys API · ServiceNow · SharePoint
  • Azure DevOps · ADP
  • Microsoft Teams / M365 · Copilot Studio
  • Reusable architecture & mentoring

Journey

Experience & education

2025 — Present

AI Engineer

Shipped 10+ production AI systems, designed a reusable agent architecture the team scales on, and helped teams adopt and integrate AI into existing workflows.

2021 — 2025

B.S. Computer Science

University of South Dakota · AI Specialization · Math Minor

Capstone: an AI Code Classifier — custom ML models that grade source-code quality from structural and logical patterns. Coursework across AI, data mining, computer architecture, and algorithms.

Beyond engineering

An engineer who also understands audiences.

For four years I ran the Barstool Sports account for the University of South Dakota — building real instincts for content, timing, and what makes things spread. The same instinct I bring to making AI tools people actually want to use.

4k → 20k+
Followers grown
4M+
Cumulative likes
5M+
Views · 3 viral posts

Work with me

How I can help

Available for contract & consulting engagements. Industry-agnostic — the same patterns translate across support, operations, internal tools, and data products.

01

Agentic Workflows

Multi-agent systems with intent routing and structured outputs that automate real, high-value workflows end to end.

02

Enterprise RAG

Knowledge assistants grounded on your documentation, with auto-sync so answers stay accurate as your content changes.

03

NL-to-Data Tools

Let non-technical teams query SQL / KQL data in plain English — with validation and execution layers that keep it safe.

04

Real-Time Automation

Concurrent, idempotent pipelines that turn manual, high-volume tasks — like live call summarization — into reliable background processes.

05

Applied ML

Gradient boosting and calibration (LightGBM, isotonic regression) for the problems where a model fits better than a prompt.

06

AI Platform Foundations

Reusable architecture, shared tooling, and patterns so your team can stand up new AI solutions quickly.

Contact

Let's talk.

Contract project, full-time role, or just a conversation about what's possible — I'd be glad to hear what you're working on.

Send an email
jensenalex3022@gmail.com(507) 430-3914LinkedInDownload résumé