01
Map the input
Understand resumes, schemas, documents, or user intent before a model touches it.
AI Engineer | Retrieval, agents, backend systems
Building Practical AI Systems That Hold Up
I build AI tools around retrieval, agent workflows, FastAPI services, and clean product interfaces. The Projects page stays connected to GitHub, so new public repos can appear without rewriting the portfolio by hand.
build flowproject map
sourceGitHub projects
pipelinesearch + ranking
serviceFastAPI endpoints
interfaceclear product UI
16public repos
7AI signals
Retrieval91%
ReviewClear
API LayerTyped
Ingestdocs
Retrievevectors
Reasonplanning
Guardchecks
ShipAPI
ATS Resume AIRAG + FastAPI
SQL CopilotText-to-SQL
Semantic RAGVector Search
recent buildATS Resume AI
Live syncRAG PipelinesEmbeddings, retrieval, grounded answers
Agent WorkflowsPlanner, tools, memory, review loops
LLM GuardrailsValidation, prompting, safer outputs
AI BackendsFastAPI contracts around model systems
ResearchCase-study notes
RAGRetrieval pipelines
AgentsPlanner and tool flows
SystemsFastAPI contracts
AI CoreProduction mindset
APIsBackend integration
LLMsPrompted applications
Build style
Clear systems before clever demos.
I like building the parts where models meet real software: the data path, the service boundary, and the interface that makes the result easy to inspect.
02
Design the workflow
Split the system into retrieval, planning, validation, and service boundaries.
03
Ship the interface
Turn the backend into a product surface people can read, test, and trust.
Toolchain
Built with the stack behind the projects.
PythonFastAPIRAGAgentsVector SearchGitHub Sync
Research notes
Featured builds with the decisions behind them.
Research Note 01
ATS Resume AI
An applied RAG system for resume analysis, ATS scoring, and job-fit feedback using local model components and retrieval.
Research question: How can an AI resume analyzer provide useful ATS feedback while keeping recommendations grounded in the user's actual resume and target role?
PythonFastAPIRAGChromaDBSentence Transformers
PythonSystem View
ParserATS RulesEmbeddingsFeedback
AI Resume IntelligenceATS Resume AI
Research Note 02
SQL Copilot
A database copilot that frames natural-language analytics as a guarded planning, SQL generation, validation, and explanation loop.
Research question: How can a natural-language SQL assistant reduce friction for analytics while making generated queries safer and more explainable?
PythonSQLLLMPrompt EngineeringValidation
PythonSystem View
PlannerSchemaSQL GuardExplain
AI CopilotSQL Copilot
System thinking
Architecture first, then model behavior.
RAG Architecture
Grounded answer flow for retrieval-first LLM applications.
User QueryQuery ExpansionHybrid SearchVector Index
Agent Architecture
Planner-tool-memory loop for multi-step intelligent automation.
UserPlannerToolsRetriever
FastAPI AI Backend
Typed API surface around model orchestration and persistence.
ClientRouteSchemaAI Service
Live GitHub corpus
Syncing16 public repositories, updated from GitHub.
New public repositories can show up in Projects automatically. The strongest builds still keep extra notes for architecture, tradeoffs, and results.
Latest repository signalATS Resume AI16Public repos
7AI systems
3Live demos
Open to AI engineering work that needs more than a demo.
I focus on the practical layer: retrieval that finds the right context, agents with understandable steps, APIs with clear contracts, and interfaces that make the work easy to review.
Retrievalgrounded context
Agentsclear tool flow
Backendtyped APIs