Building production-oriented AI systems for reasoning, retrieval, automation, and intelligent decision support.
I am an NLP Developer and AI Research Engineer focused on turning AI concepts into useful, testable, and maintainable systems.
My work spans the complete engineering lifecycle—from data collection, ETL, validation, and feature pipelines to semantic retrieval, agentic workflows, backend APIs, deployment, and production monitoring.
I am especially interested in:
- Deterministic and explainable AI for grounding probabilistic language models
- RAG, semantic search, and structured memory for knowledge-intensive applications
- Agentic systems with tools, state, evaluation, and observable execution
- ML and MLOps platforms that connect experiments to dependable services
- Systems engineering through a Rust-first operating-system project
A Rust-first x86_64 operating system developed from the boot layer upward.
The project currently explores UEFI boot, architecture-specific initialization, hardware discovery, PCI enumeration, reproducible builds, and an incremental kernel roadmap. It demonstrates low-level engineering, careful interface design, documentation, and long-horizon system development.
Rust UEFI x86_64 PCI Systems Engineering
A deterministic symbolic-reasoning and structured-memory layer for grounded, auditable LLM applications.
SanTOK investigates how knowledge graphs, explicit inference rules, constraints, memory, and inspectable reasoning traces can complement probabilistic language models. The goal is to reduce unsupported generation by grounding supported responses in traceable facts and bounded reasoning processes.
Python Symbolic AI Knowledge Graphs LLM Grounding Explainable AI
A reproducible research framework for studying the emergence of computational structure and identity.
Threshold Onset explores how observable structure can emerge through state transitions, action, trace, and repetition before higher-level symbolic interpretation. The repository emphasizes deterministic experimentation, reproducibility, and explicit research documentation.
Python Reproducible Research Deterministic Systems Structure Discovery
| Project | Focus |
|---|---|
| FinTrack | TypeScript financial-management application under active development |
| newdaily | Cross-platform TypeScript utility suite with an Android build workflow |
| universal_tts_system | Modular Python text-to-speech experimentation and application integration |
| Job_Automation | Python automation workflows and supporting tooling |
| Area | What I work on |
|---|---|
| Generative AI & NLP | RAG, semantic retrieval, embeddings, structured memory, LLM integration, prompt and retrieval evaluation |
| Agentic AI | Tool-using agents, stateful workflows, orchestration, multi-agent systems, guardrails, observability |
| Machine Learning | Data understanding, EDA, feature engineering, model training, evaluation, error analysis |
| MLOps | Experiment tracking, data and model versioning, registries, CI/CD, deployment, monitoring, drift detection |
| Data Engineering | ETL pipelines, data validation, business-rule mapping, cloud storage workflows, production troubleshooting |
| Backend Engineering | FastAPI services, REST APIs, asynchronous workflows, persistence, containerized applications |
| Systems Engineering | Rust, UEFI, x86_64 architecture, hardware discovery, low-level interfaces |
- Consolidating the SanTOK ecosystem into one coherent, evaluated AI platform
- Extending OperatingSystem through staged hardware and kernel development
- Developing an Intelligent Predictive Maintenance & GenAI Operations Platform
- Strengthening automated evaluation, testing, observability, and CI/CD across AI projects
- Converting research ideas into reproducible experiments and versioned releases
Understand the problem
→ validate the data
→ establish a reproducible baseline
→ design the system
→ evaluate failure modes
→ automate delivery
→ observe production behaviour
→ improve with evidence
I value honest evaluation, explicit trade-offs, clear documentation, reproducible results, and systems that remain understandable after the first demo.
I am interested in collaborating on:
- Production RAG and semantic-search systems
- LLM evaluation and agent reliability
- Explainable and deterministic AI
- ML/MLOps platforms and developer tooling
- Rust systems and operating-system development
- Open-source projects with measurable user or engineering impact