Backend Engineering | Applied AI | Enterprise Applications
I'm an Information Technology Analyst with 6+ years of software-development experience. I work on enterprise backends and build practical AI tools, combining Python, .NET, Java, and Azure with hands-on experiments in language models and machine learning.
My focus is on useful systems: clear workflows, traceable evidence, reliable APIs, and human control where it matters.
For the clearest view of my current work, begin with RecallOps, an evidence-focused incident memory assistant, or Local-First Coding Assistant, a workspace-aware coding tool with approval-gated actions.
An incident-response assistant that retrieves relevant resolved incidents and presents the supporting evidence for an operator to review. Built around FastAPI, durable memory, provenance, and human approval rather than automatic remediation.
Status: Local workflow implemented; cloud integrations and hosting still need live validation.
A coding-assistant project using FastAPI, Angular, SQLite, and Ollama. It includes streaming conversations, repository exploration, local search, and Git context, with approval-gated file writes and terminal commands.
Focus: Local models, workspace-aware tools, and explicit control over actions.
An MVP for small-business lead follow-up using Flutter, PHP, MySQL, and Gemini. It combines lead management, reminders, and AI-drafted replies in a workflow where the user reviews and sends each message.
Focus: Practical business automation without automatic outbound messaging.
A Kaggle multiclass-classification project with reproducible experiments, stratified validation, and explicit acceptance criteria for model changes.
Documented CatBoost baseline: 0.94909 out-of-fold balanced accuracy and 0.94979 public leaderboard score. These are separate evaluation results, not interchangeable accuracy figures.
A personal speech-synthesis project using Coqui TTS and VITS, covering audio preparation, dataset quality checks, GPU training, checkpointing, and inference.
Focus: Hands-on model-training and inference workflows.
| Area | Technologies |
|---|---|
| Backend engineering | Python, FastAPI, C#, ASP.NET Core, Java, Spring Boot |
| Data and infrastructure | SQL, MongoDB, SQLite, Azure, Azure Functions, queues |
| Applied AI and ML | LLM applications, RAG, Ollama, PyTorch, CatBoost, speech synthesis |
| Application development | Angular, REST APIs, authentication, role-based access control |
Grounded AI assistants, agent memory, local-model workflows, and reproducible evaluation - bringing backend-engineering discipline to AI applications.
Understand the problem -> build a useful baseline -> test the failure cases -> document the evidence -> improve deliberately.
I value reproducible results and clear limitations over inflated claims. Each repository is the source of truth for its setup, scope, and implementation status.
Interested in collaborating on practical AI tools, backend systems, or reproducible ML projects? Explore all repositories or reach me on LinkedIn.