AI Engineer & UiPath Tech Lead building and maintaining enterprise automations in production, combining UiPath with Python, Playwright, REST APIs, and AI agents.
Over the years as an RPA developer and tech lead, most of what I have learned comes from keeping robots reliable in real production conditionsβwhen web portals change layout, inputs arrive in unexpected formats, or business teams depend on unattended queues finishing on time.
My engineering approach: use deterministic automation where rules are clear, and introduce AI where interpretation, reasoning, or flexibility adds practical value. Keep the AI step focused on the part that needs judgment, and let Python or UiPath execute the rest.
1. NAVIA_PATH β Open-Source AI Browser Agent
Prompt-driven browser automation for UiPath and Windows workstations, powered by Python, Playwright, and multi-provider LLMs.
- 5-Step Agentic Loop: Natural-language prompt β Targeted browser observation β LLM decides next action β Playwright executes action β Observes new state until completion β Returns structured JSON.
- UiPath Studio Activities: Native
NAVIA PATH - ChromeandNAVIA PATH - Edgeactivities returningout_ExecutionResult. - Standalone Windows Executable: Ships with compiled Windows runtime (
NAVIA_PATH.exe) and environment validation (CHECK_ENVIRONMENT), requiring zero manual Python setup on target machines. - Model Agnostic: Works with Google Vertex AI, Gemini Direct, OpenAI, Anthropic Claude, and Azure OpenAI.
π Explore NAVIA_PATH Repository
2. support-request-AI-agent-UiPath β UiPath + Generative AI Triage
Practical division of labor between Generative AI intent classification and deterministic UiPath RPA.
- Separation of Concerns: Generative AI classifies unstructured customer intent from incoming JSON tickets (
Deposit,Withdraw, orProduct) via constrained system prompt engineering. - Deterministic Orchestration: UiPath Studio workflows and a BPMN 2.0 flow remain in full control of queues, file routing, and error handling.
- Production Architecture: File ingestion monitoring, Newtonsoft JSON deserialization, modular dispatcher/performer workflow design.
π Explore support-request-AI-agent-UiPath
| Category | Technologies & Tools |
|---|---|
| Enterprise RPA & Orchestration | UiPath Studio, Orchestrator, REFramework, Dispatcher/Performer, UiPath Maestro, Document Understanding, Automation Anywhere |
| Web & Scripted Automation | Python, Playwright, REST APIs, C#, VB.NET, PowerShell, SQL / SQLite |
| Agentic AI & LLMs | AI Agents, Tool Calling, Structured JSON Outputs, Prompt Engineering (Vertex AI, Gemini, OpenAI, Claude, Azure OpenAI) |
| Delivery & Engineering | Git, GitLab, GitHub, CI/CD, BPMN 2.0, Jira, Scrum / Agile |
- UiPath Certified Professional Agentic Automation Associate β UiPath (Apr 2026)
- CS50's Introduction to Artificial Intelligence with Python (CS50 AI) β Harvard University (Oct 2026)
- Advanced Automation Certification β Automation Anywhere (Jan 2026)
- CS50x: Introduction to Computer Science β Harvard University (Jun 2024)
- SS&C Blue Prism Robotic Process Automation Professional Certificate β SS&C Blue Prism (May 2024)
- Atlassian Agile Project Management Professional Certificate β Atlassian (Feb 2025)
- Scrum: Advanced β LinkedIn Learning / PMI (Mar 2025)
- McKinsey Forward Program β McKinsey & Company (Mar 2024)
- LinkedIn: linkedin.com/in/berjamy
- Primary Article: AI Engineer from UiPath RPA to Agents
- Upwork: upwork.com/freelancers/~0130438826692ba99b
- YouTube: youtube.com/@imad_ui
- Email: imad.berjamy@outlook.com
