[Caice-csse] A PhD applicant that looks very good on paper

Sun May 25 09:00:36 CDT 2025

All,
This person looks really good on paper: BS in CS from a very good Indian university, MS in DSE from one of the top 10 engineering schools in India, very high BS and MS GPAs (by grading standards in top Indian schools), nearly 10 years of post-BS and post-MS experience in some of the top IT companies in India, and seems to have attended (remotely I suppose) AI/ML/DS summer schools at top universities – U Chicago, Duke, NYU, and Oxford, and claims to have attended ICML and NAACL on travel scholarships. Either this is all too good to be true, or he is a terrific potential PhD applicant that one of you may want to snap up ☺
FYI,
- Hari

From: ParthaPratim Saha <technical.partha at gmail.com>
Sent: Sunday, May 25, 2025 8:02 AM
To: N Narayanan <naraynh at auburn.edu>
Subject: Prospective PhD in your Lab

You don't often get email from technical.partha at gmail.com<mailto:technical.partha at gmail.com>. Learn why this is important<https://aka.ms/LearnAboutSenderIdentification>
Dear Prof Hari

I am Partha Pratim Saha, interested in pursuing a PhD in AI, Natural Language Processing(NLP) & Large Language Models (LLMs) under your supervision. I completed my Master's Degree with Distinction in Data Science and Engineering from Birla Institute of Technology and Science (BITS), Pilani in India, with an emphasis on Machine Learning, Deep Learning, interpretable NLP & LLMs, Retrieval Augmented Generation (RAG), and AI Agentic workflow areas. I am interested in working on automatic and modularized toolbox for the Knowledge Discovery project.

I developed a conversational dialogue system and worked on a medical search engine using databases, text documents, and images. I worked on Open-GPTx, GPT-o1, T5, GPT-4, and Llama via HuggingFace Hub to develop Multihop & Multiquery Retrieval Augmented Generation (RAG) pipeline with Llama Index, Vector and BM25 retriever. I have an advanced Machine Learning, Data Science, Maths, & Statistics and NLP background.

I worked on the safety of Agentic AI tools. I explored hybrid models combining foundation LLMs (e.g., Llama 3, Mixtral 8) with RAG pipelines using vector databases (e.g, Pinecone) and used Multi-agent frameworks, e.g., LangChain Agents, AutoGen. I have extensive experience with Python, PyTorch, and TensorFlow.

I have explored how Chain-of-thought, Tree-of-Thought, Chain-of-Knowledge, and Chain-of-Verification could be integrated end-to-end to develop better logic, reasoning, and fact-checking abilities in LLMs. To see these aspects, I did thought simulations and experimented with prompts in Mixtral-8x7B, Llama2-70B, GPT-4, and Gemini Pro LLM.

I focused on LLM-as-a-judge. I am passionate about researching and making ethical and responsible AI accessible to everyone and open-sourcing it. Towards this work, as a Google's Udacity Bertelsmann Technology scholar, I got accepted into the 2024 AI+Science Summer School at the University of Chicago<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F8f81f76f.streaklinks.com%2FCcEVpoUX1xK3gJRxcggisvfa%2Fhttps%253A%252F%252Fdatascience.uchicago.edu%252Fevents%252Fai-science-summer-school-2024%252F&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7Ca5a4cd0c18b74b4e4af508dd9b948613%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638837784404745699%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=kdihPmenxUy%2FirROQEI1Her%2B7UGVQNZGtwCdtd5io5s%3D&reserved=0>, New York University's (NYU) AI Summer School<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F8f81f76f.streaklinks.com%2FCcEVpoQFxW7idCjbSALySeK7%2Fhttps%253A%252F%252Fcims.nyu.edu%252Fai%252Feducational-programs%252Fpathways-ai%252F&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7Ca5a4cd0c18b74b4e4af508dd9b948613%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638837784404770838%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=ZT%2FgqXQ%2BvEh0do6iouirIDe%2BuSzQnOEatY9m8%2B3rBOM%3D&reserved=0>'22, and was awarded to attend ICML'21, CLPsych for NAACL’22 & ACL’22 conferences.

I have experience in research proposal writing. One of my research proposals on NLP was selected by NeuroMatch Academy<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F8f81f76f.streaklinks.com%2FCcEVpoYLS4BTJ38H_gkwSbJP%2Fhttps%253A%252F%252Fdeeplearning.neuromatch.io%252Ftutorials%252Fintro.html&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7Ca5a4cd0c18b74b4e4af508dd9b948613%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638837784404785074%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=loOaUZI%2Fh0yeFwkAIeYlNAqGB3H5Xr0f95t9LAetnl0%3D&reserved=0>, and thereafter, I was selected as a scholar in the Impact Scholar Program(ISP)<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F8f81f76f.streaklinks.com%2FCcEVpoU-vlQu3D2l3AfIX3m6%2Fhttps%253A%252F%252Fneuromatch.io%252Fneuromatch-impact-scholars-program%252F&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7Ca5a4cd0c18b74b4e4af508dd9b948613%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638837784404799820%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=waTRWwrEy1%2BsMzLJLLlXswC%2F8MC6W%2FkX26nhvn%2Bgpx8%3D&reserved=0>. During my post-graduate research, I worked with Prof. Miad Faezipour<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F8f81f76f.streaklinks.com%2FCcEVpoYNTQHBM06MJAJ6OzYX%2Fhttps%253A%252F%252Fscholar.google.com%252Fcitations%253Fhl%253Den%2526user%253D2NvRZn4AAAAJ%2526view_op%253Dlist_works%2526sortby%253Dpubdate&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7Ca5a4cd0c18b74b4e4af508dd9b948613%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638837784404813857%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=Dvt0tnHzrxaVMgi2AQh2ZPJ4tNz84Vu%2Bs37k2IN6pYE%3D&reserved=0> from Purdue University, USA and Dr. Naresh Sehgal<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F8f81f76f.streaklinks.com%2FCcEVpoYV1TCdds5sqQ0yn6p7%2Fhttps%253A%252F%252Fscholar.google.com%252Fcitations%253Fhl%253Den%2526user%253D7DhuLLMAAAAJ%2526view_op%253Dlist_works%2526sortby%253Dpubdate&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7Ca5a4cd0c18b74b4e4af508dd9b948613%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638837784404827525%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=gisKGQWpwgXcpEJrYWbl7QTcdEB6NG2jdsNpXP396JI%3D&reserved=0>, my industry supervisor, on a personalized GPT-like system using Federated Learning for healthcare and wrote a paper<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F8f81f76f.streaklinks.com%2FCcEVpoYGRodz0ZddtgWz4nb6%2Fhttps%253A%252F%252Fdrive.google.com%252Ffile%252Fd%252F17A9xwEm4JevpemImidTzkU1_bcwcKArL%252Fview&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7Ca5a4cd0c18b74b4e4af508dd9b948613%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638837784404841240%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=WJ6%2F6pCIIgDnNagdfD1RzHsFuZZLYsVq4Cx%2FWhRN1e4%3D&reserved=0> which got published this year at a conference<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F8f81f76f.streaklinks.com%2FCcEVpoUSkgakU8aYRgBNe7J8%2Fhttps%253A%252F%252Famerican-cse.org%252Findex.html%252F&data=05%7C02%7Ccaice-csse%40eng.auburn.edu%7Ca5a4cd0c18b74b4e4af508dd9b948613%7Cccb6deedbd294b388979d72780f62d3b%7C0%7C0%7C638837784404859435%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=ki195KBiAHT8WUA%2BlyBIMWA1lvjqGwaxdlNqIg%2BcZ5Y%3D&reserved=0> in Las Vegas, USA.

I look forward to the possibility of contributing to your research group to pursue my PhD at your lab. I have attached my CV for your kind reference. Thank you.
Sincerely,
Partha Pratim Saha
[https://mailfoogae.appspot.com/t?sender=adGVjaG5pY2FsLnBhcnRoYUBnbWFpbC5jb20%3D&type=zerocontent&guid=a2dabbb6-da24-457f-bd3c-d880d5dd5bde]ᐧ
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.eng.auburn.edu/pipermail/caice-csse/attachments/20250525/c6937dec/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: CV_Partha_2025.pdf
Type: application/pdf
Size: 256451 bytes
Desc: CV_Partha_2025.pdf
URL: <http://mailman.eng.auburn.edu/pipermail/caice-csse/attachments/20250525/c6937dec/attachment-0001.pdf>