AI / ML ENGINEER

Building production-oriented AI systems.

I design and deploy Generative AI, Agentic AI, RAG, and LLM applications with strong engineering controls around validation, evaluation, evidence, observability, and reliability.

Explore projects ↓Try live demo →GitHub ↗

system profileonline

Probabilistic intelligence is paired with deterministic controls — typed outputs, evidence checks, stateful workflows, quality gates, and regression tests.

{
  "focus": [
    "Agentic AI",
    "RAG",
    "LLM Systems",
    "AI Automation"
  ],
  "default": "engineered, not improvised"
}

Selected work

Systems I built to be trusted.

Five portfolio projects across modernization, evidence, procurement, automation, and browser intelligence.

01Deployed

Agentic software modernization

LegacyLens

Analyze legacy repositories, plan migrations, transform code, and validate behavior with automated quality gates.

Agentic AICode IntelligenceAgnoFastAPIDocker

Explore project

02Deployed

Verified RAG & research

EvidenceFlow

Hybrid retrieval, reranking, web research, evidence verification, and fail-closed answer generation.

LangGraphQdrantBM25RRFTavily

Explore project

03Deployed

Procurement intelligence

QuoteSense

Turn messy quotations into validated, comparable procurement intelligence with deterministic scoring and review.

Document AIPydanticFastAPIProcurement

Explore project

04Deployed

Agentic automation

FlowPilot

Asynchronous AI workflows with approvals, evidence, persistence, and observability.

LangGraphFastAPIRedis/RQPostgreSQLObservability

Explore project

05Deployed

AI-assisted web QA

WebQA Intelligence

Discover dynamic web interactions, score QA risk, generate evidence-grounded tests, and compare regressions.

PlaywrightBrowser AutomationQAAgents

Explore project

Interactive demo · BYOK

Try the shared inference boundary.

Bring your own provider key to create a temporary session. The portfolio never stores the provider credential.

Interactive demo · BYOK

Try the interactive demo

Use your own provider key to open a short-lived inference session. The credential is used only for the session request.

Provider

Model ID

Provider API keyUsed once to create the temporary portfolio session. It is not stored by the portfolio.

One session · reusable across project pages · provider key never stored

Engineering approach

LLMs are components, not the application.

The architecture keeps probabilistic reasoning close to deterministic interfaces, evidence, state, validation, and deployment controls.

01Unstructured data

02Retrieval / processing

03LLM / agents

04Structured validation

05Deterministic logic

06Evidence / quality gates

07Evaluation / observability

08Deployment / API