Skip to content
View TomaIjatomi's full-sized avatar
6 followers · 14 following

Block or report TomaIjatomi

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
TomaIjatomi/README.md

Toma Ijatomi

AI & Data Practitioner
Building at the intersection of agentic AI, safety, and responsible deployment.

Based in Dublin, Ireland. Working on RAG pipelines, agentic systems, and AI evaluation frameworks.
Program Lead, Women in AI Ireland · Facilitator, AI Safety Dublin.

🌐 tomaijatomi.com · 📺 YouTube · 💼 LinkedIn

What I'm Working On

  • FinSheet Agent - A multi-agent system for financial spreadsheet question answering.
  • Arxiv Source-Grounding Assistant - A production-grade RAG system, built eval-first, that checks claims against academic AI papers. (LlamaIndex, FastAPI, PostgreSQL, OpenSearch, Redis, Airflow)
  • TraceBench - Task-level Reasoning and Agentic Cognition Evaluation (TRACE) Benchmark for evaluating Self-Monitoring in Multi-Step Agent Tasks
  • Gravitas — Leadership communication app with AI coaching and voice synthesis (Next.js, Google Cloud TTS)
  • Exploring adversarial AI, prompt injection defence, and security patterns for agentic systems

Selected Talks & Workshops

Let's Connect

Open to collaborations, speaking invitations, and interesting AI problems.
Contact me on LinkedIn or find me at tomaijatomi.com.

Pinned Loading

  1. finsheet-agent finsheet-agent Public

    Multi-agent system for financial spreadsheet QA. CoDaS architectural pattern via MCP. 100% on hard tier, 97% stratified.

    Python 1

  2. arxiv-grounding-assistant arxiv-grounding-assistant Public

    Eval-first RAG that checks claims against arXiv papers; cites the spans it uses, abstains when the corpus is silent, and measures every component against a frozen, passage-anchored gold set.

    Python

  3. talks-and-workshops talks-and-workshops Public

    Jupyter Notebook 1

  4. youtube youtube Public

    Practical AI engineering for production - code, demos and reseources from my YouTube channel covering agentic AI security, evals, and reliability.

    Python

  5. robust-malware-detection-ensemble robust-malware-detection-ensemble Public

    Stacked ensemble malware detection with adversarial (MalGAN-inspired) robustness evaluation using EMBER dataset.

  6. multi-output-product-classification multi-output-product-classification Public

    Multi-output LSTM-based text classification for large-scale e-commerce product categorization under class imbalance.

    Jupyter Notebook