🟢 Open to work
A computer science sophomore, passionate about building AI systems. From engineering RAG pipelines with Pfizer to developing AI automation processes at Peck Hannaford + Briggs
A collection of projects from my academic work, personal projects, and Pfizer externship—showcasing my skills as I continue to grow.
.jpg)
Pfizer · Aug 2026 · ✅ Verified by Extern
Prototype AI-powered document intelligence with Pfizer—using OCR, LLMs, and RAG to automate real enterprise PDF workflows and build a standout portfolio project.
Built and deployed a full-stack healthcare reporting application integrating the CMS Provider Data Catalog API through a FastAPI proxy to retrieve facility details by CMS Certification Number (CCN), including address, bed count, and star ratings. Implemented branded PDF and Word report generation, combining live CMS data with operational notes; deployed the React frontend on Vercel and FastAPI backend on Render, addressing CORS restrictions and free-tier cold-start latency.
Co-built and programmed an autonomous EV3 robot in LabVIEW to sort shredded disaster-relief materials, using weight detection to distinguish organic from metal waste and route bins to designated processing stations; project specifications included 90% sorting accuracy and a 120-second maximum path time. Integrated 4 sensors (2 color sensors for closed-loop navigation and 2 touch sensors for bin handling) with a motorized arm to automate bin pickup, placement, delivery, and return-to-station operations across terrain variations of up to 2 cm. Led documentation for a 3-person engineering team, maintaining the design notebook, Gantt chart, and meeting agendas while authoring project deliverables across development.
A computer science sophomore, passionate about building AI systems. From engineering RAG pipelines with Pfizer to developing AI automation processes at Peck Hannaford + Briggs
I am Mahi Sheth, a 2nd-year computer science student at the University of Cincinnati pursuing a B.S. focused on AI. I’ve built full-stack projects and an autonomous robot and completed an AI externship with Pfizer. I am currently a software engineering intern and open to roles that strengthen my resume and grow my skills in Python, web stacks, LLMs, and cloud tooling.
Externships
Pfizer Advanced: AI-Powered Document Insights & Data Extraction Externship
Completed · Aug 2026
Experience
Software Engineering Co-op
Peck Hannaford + Briggs · August 2026 – Present
Education
University of Cincinnati
B.S. Computer Science — College of Engineering & Applied Sciences · Class of 2030
William Mason High School
High School Diploma · Class of 2025
Skills
✅ Verified by Extern
Prototype AI-powered document intelligence with Pfizer—using OCR, LLMs, and RAG to automate real enterprise PDF workflows and build a standout portfolio project.
The work prototyped AI-powered document intelligence for pharmaceutical PDFs, producing technical notes, runnable Colab notebooks, extraction scripts, OCR comparisons, and RAG experiment logs. Deliverables included a transformer technical summary, preprocessing and PyMuPDF notebooks with layout analysis, a three-way OCR evaluation, RAG pipeline implementations with recorded retrievals, and a JSON
I delivered a coherent artifact set: a transformer/LLM technical brief, Colab notebooks for cleaning and image enhancement, PyMuPDF extraction with bounding boxes and parsing notes, a Tesseract/PaddleOCR/EasyOCR comparison, RAG experiment logs for three configurations, and a JSON mapping of split documents with doc_id and doc_type.
Built and deployed a full-stack healthcare reporting application integrating the CMS Provider Data Catalog API through a FastAPI proxy to retrieve facility details by CMS Certification Number (CCN), including address, bed count, and star ratings. Implemented branded PDF and Word report generation, combining live CMS data with operational notes; deployed the React frontend on Vercel and FastAPI backend on Render, addressing CORS restrictions and free-tier cold-start latency.
Co-built and programmed an autonomous EV3 robot in LabVIEW to sort shredded disaster-relief materials, using weight detection to distinguish organic from metal waste and route bins to designated processing stations; project specifications included 90% sorting accuracy and a 120-second maximum path time. Integrated 4 sensors (2 color sensors for closed-loop navigation and 2 touch sensors for bin handling) with a motorized arm to automate bin pickup, placement, delivery, and return-to-station operations across terrain variations of up to 2 cm. Led documentation for a 3-person engineering team, maintaining the design notebook, Gantt chart, and meeting agendas while authoring project deliverables across development.
