Iljard Xhani · AI Engineering Download CV

AI Application Engineer · LLM + RAG shipped with eval gates

I Ship Reliable LLM Features — With Measurable Quality

6 end-to-end case studies with architecture, eval results, and deployment evidence (logs, traces, screenshots).

6 case studies Architecture diagrams Eval reports Trace screenshots Deploy notes
View Case Studies Contact
  • Production RAG: ingestion pipelines, retrieval tuning, citations, and session memory.
  • Evaluation Discipline: faithfulness checks, regression gates, and trace-based debugging.
  • Product Engineering: FastAPI services, auth boundaries, rate limits, and auditability.

Hiring Snapshot

What Teams Can Expect in the First 90 Days.

I focus on fast alignment, measurable quality, and production-safe execution.

First 30 Days

Audit the current AI workflow, define success metrics, and establish an implementation plan tied to user and business outcomes.

Architecture Audit KPI Baseline Risk Register Delivery Plan

By Day 90

Ship reliable AI features with retrieval quality checks, runtime guardrails, and clear observability for continued iteration.

Production Workflow Eval Gates Monitoring Iteration Playbook
Read Full Profile

Capabilities Snapshot

Core Areas for AI Application Engineering.

RAG Architecture

Multi-source ingestion, chunk strategy, retrieval tuning, and citation-grounded answer flows.

Pipeline Design · Retrieval Quality

LLM App Layer

Prompt routing, function/tool calling, schema-constrained outputs, and deterministic APIs.

Prompting · Orchestration · APIs

Evaluation + Ops

Offline/online eval loops, observability, release checks, and rollback-safe deployment flow.

Evals · Monitoring · Reliability

Role Fit Guide

Which Projects to Send for Different Company Types.

Use this map to tailor each application in under a minute.

Internal Knowledge and Productivity Teams

Start with DocChat RAG and Knowledge Base Builder to demonstrate grounded answers and scalable ingestion.

Knowledge Systems

Open Full Project Gallery

Projects Snapshot

Case Studies With Technical and Delivery Evidence.

Each case study follows one pattern: Problem, Architecture, Evidence, and Delivery decisions.

DocChat RAG (Production)

RAG Production

Grounded internal knowledge assistant with citations, retrieval confidence, and session memory.

Evidence: citation checks, fallback-rate monitoring, and p95 latency tracking.

Open case study →

RAG Evaluation Lab

Evaluation Regression

Automated faithfulness, relevance, and regression scoring used as release gates.

Configured gates: faithfulness >= 0.88 and p95 latency <= 1900ms.

Open case study →

Mini SaaS: RAG for Teams

SaaS Featured

Multi-tenant AI app surface with auth, org isolation, billing flow, rate limits, and audit logs.

Validation: tenant-isolation tests and per-plan burst rate-limit checks.

Open case study →
See Full Gallery (6 Projects)

Process Snapshot

Delivery Process I Use for AI Features.

1. Scope and Retrieval Strategy

Context Model · Data Model · Success Metric

Define user questions, source-of-truth data, and quality targets before model tuning.

2. Build LLM + Tooling Layer

Prompt Contracts · Tool Calls · Error Paths

Implement deterministic interfaces and guardrails so behavior stays predictable.

3. Evaluate, Observe, Iterate

Eval Runs · Tracing · Release Gate

Compare outputs, inspect failures, and tune retrieval/orchestration before release.

Open Full Process Page

Contact Snapshot

Open to AI Application Engineer Interviews.

If your team is building AI features that must be stable, measurable, and fast, I can help.

Send the role scope, your current stack constraints, and what you need to ship or improve in the first 90 days, and I will reply with how I would approach it in practice, where I can help fastest, and realistic first steps within about a day.

Compose in Gmail View Resume (PDF)
Compose in Gmail LinkedIn Full Contact Page