Artificial Intelligence Researcher · Factiverse
Remote · Oct 2023 — Dec 2025
- Research
- NLP
- PyTorch
- LangChain
- vLLM
- HuggingFace
- Mechanistic Interpretability
- Fact-Checking
Worked on multiple research automated fact checking research projects. I've published research papers on question generation and information retrieval to CIKM and ICTIR.
I worked on multiple research projects as part of my internships including QuestGen and a paper on sparse retrieval. Through this experience, I've been able to get hands on experience using linux servers, nvidia GPUs, and more importantly, experience training and finetuning LLMs.
I got hands-on experience using pytorch, langchains, vllm and other AI related libraries on SoTA (state of the art) LLMs.
As apart of my QuestGen Project (which is listed under publications and has its own page here), I finetuned models on five datasets, ran jobs to train models in multi-day long jobs, and ran my own experiements to evaluate the effectiveness of question generation for fact checking with LLMs.
I also wrote another research paper on the effectiveness of different IR methods for fact checking with LLMs. For this paper, I ran ablation studies on SoTA fact checking pipelines in order to identify the the most important step of the automated fact checking process.
Outside of research, I worked on building the following systems for automated fact checking
- I built a multi-modal fact-checking system that verifies claims across media formats, folding visual evidence into textual claims.
- I worked on developing a speaker diarization pipeline to enable real-time automated fact checking of political debates.