GSoC 2026 Interest: ML-assisted Anonymization Layer for Greek Datasets

[GovindhKishore]

Hi everyone!

I’m Govindh Kishore, a Mathematics and Computing student at BIT Mesra. I am very interested in contributing to the ML-assisted Anonymization Layer for GlossAPI.

I specialize in building search and retrieval pipelines that handle complex, unstructured data. My relevant technical background includes:

Advanced Search & Indexing: Built a Semantic Code Search Engine using Python AST for structural analysis.

RAG & NLP Pipelines: Developed an Assessment Recommendation System using Sentence-Transformers for embeddings and a two-stage retrieval process (candidate generation + LLM reranking) served via FastAPI.

Data Engineering: Experienced in handling dirty data and OCR noise from building large-scale web scrapers (BeautifulSoup).

Regarding the project:
Since GlossAPI handles diverse Greek datasets, I believe my experience with PyLucene will be useful for building efficient lookups for rule-based anonymization (e.g., blacklists of sensitive organizations).

Questions for the mentors (@nikostsekos @myrsiniioannou @jimmmyss):

I don't speak Greek, but I am proficient in working with Unicode/UTF-8 and utilizing pre-trained NLP models. Will this be an issue for the ML-assisted part of the project?

Are there any specific issues or "warm-up" tasks in the glossAPI repo related to data preprocessing or regex filtering where I could start contributing?

I’ve explored the repository and would love to start working on a small Proof of Concept (PoC) for the anonymization module.

Best regards,
Govindh Kishore
GitHub Profile