CIS 6930 Spring 26

Logo

Data Engineering at the University of Florida

Presentation

Due: Monday, April 20, 2026 by 8:30 AM (slides + demo recording in repo) Points: 150 Format: 5-7 minute group presentation with recorded demo

Overview

The final presentation showcases your project in a small-group setting. On Monday, you present to a group of 4-5 classmates working on similar topics. Your group votes on the best presentation. On Wednesday, group winners present to the whole class and the class votes on the overall best. Winners earn extra credit.

Schedule

Day What Happens
Mon, Apr 20 Small-group presentations (5 groups, simultaneous)
Mon, Apr 20 Each group votes on the best presentation
Wed, Apr 22 Group winners present to the whole class
Wed, Apr 22 Class votes on overall best, extra credit awarded

Extra Credit

Group Assignments

View the full groups page — formatted for display in class.

Groups are organized by project topic so that presenters and audience share enough context to ask good questions and give informed feedback.

Group A — Self-Healing and Adaptive ETL

Student Project
Atul Arun Self-Healing Civic Data Pipelines with MCP-Orchestrated Inspection
Nikhitha Nagabhyru Self-Healing ETL Pipelines via LLM Orchestration and RAG
Sai Meghana Barla Drift-Aware Self-Healing ETL Framework
Vivek Chenganassery Adaptive Context Compression for Large Log Datasets
Sai Teja Appani Benchmarking LLM-Orchestrated vs Traditional ETL

Group B — MCP-Orchestrated Data Integration

Student Project
Harris Barton LLM-Orchestrated Data Integration from Heterogeneous Book APIs
Kevin Tran LLM-Orchestrated Gainesville Open Data Integration Pipeline
Vatsal Shah Cross-City Building Permit Integration with MCP-Based ETL
Vittal Chintamaneni NavFusion: LLM-Based Route Optimization Using MCP
Siyuan Pan Deterministic ETL Pipeline for NYC 311 and Restaurant Data

Group C — LLM vs Traditional Methods

Student Project
Ian Arnold LLM vs Regex for Clinical Note Extraction
Sri Ashritha Appalchity LLM vs Trained Classifier for Entity Resolution
Sanya Chaturvedi Comparing Rule-Based and LLM-Orchestrated Pipelines
Zachary Zeng Comparing Multi-LLM Agents: Decomposition vs Augmentation
Xiaomeng Xiong Language Intervention Framework for Evaluating LLMs

Group D — Domain-Specific LLM Pipelines

Student Project
Rukaiya Khan F1 Race Strategy Intelligence Pipeline
Sanjeev Kamath Travel Risk Assessment Stability Analysis
Palavalli Shyam LLM-Orchestrated Data Pipeline for Job Market Extraction
Kanishka Dhaundiyal LLM-Driven Session Prediction with RAG
Zachary Allen RAG-Augmented Search for PubChem Database

Group E — Security, Cost, and Specialized Pipelines

Student Project
Adnan Farid LLM-Orchestrated Data Cleaning with MCP Tools
Juan Veliz LLM-Augmented Security Triage Pipeline over GitHub Code
Shane Thomas Agentic Knowledge Graphs for Lateral Movement Detection
Jiangwei Wang Cost-Aware Hybrid LLM Pipeline for Municipal Permit Data

Presentation Structure (5-7 minutes)

Section Time Content
Introduction 1 min Problem, motivation, research question
Approach 1-2 min System architecture and key design decisions
Demo Recording 2-3 min Pre-recorded pipeline walkthrough (see below)
Results 1-2 min Key findings with evidence
Takeaway 30 sec One sentence the audience should remember

What to Cover

Introduction (1 minute)

Approach (1-2 minutes)

Demo Recording (2-3 minutes)

Results (1-2 minutes)

Takeaway (30 seconds)

Recorded Demo (Required)

Every presenter must include a pre-recorded screen recording of their pipeline in action. You do not need to run code live. Step through the pipeline and show what it does.

What to Record

Recording Tips

Submission

Place the recording in your project repo:

cis6930sp26-project/
├── presentation/
│   ├── slides.pdf
│   ├── demo.mp4          ← screen recording
│   └── demo_notes.md     ← (optional) notes for your narration
└── ...

If the file is too large for GitHub, upload to YouTube (unlisted) or Google Drive and put the link in demo_notes.md.

Monday In-Class Schedule

Time Activity
8:30-8:35 Instructions, move to group stations
8:35-9:10 Group presentations (5 students x ~6 min)
9:10-9:15 Vote for best presentation in your group
9:15-9:20 Results, announce Wednesday schedule

Presentation order within each group: alphabetical by last name.

Each group will occupy a section of the room. Present from your laptop screen. Group-mates gather around to watch. A 7-minute hard cap applies; I will circulate as timekeeper.

Wednesday In-Class Schedule

Time Activity
8:30-8:33 Announce the 5 group winners
8:33-9:08 Winners present to whole class (7 min each, with brief Q&A)
9:08-9:13 Class votes for overall best
9:13-9:20 Results, extra credit, course wrap-up

Rubric (150 points)

Criterion Weight Points What Evaluators Look For
Content 30% 45 Clear problem, approach, results, and takeaway
Clarity 23% 35 Logical flow, audience can follow without prior context
Demo Recording 27% 40 Pipeline walkthrough is clear, narrated, shows key steps
Delivery 20% 30 Prepared, audible, engages the audience, within time
Total 100% 150  

Scoring Scale

Score Meaning
5 Excellent — conference-quality presentation
4 Good — professional and engaging
3 Satisfactory — gets the message across
2 Below average — hard to follow or missing elements
1 Poor — does not meet basic requirements

Common Mistakes to Avoid

  1. Too much text on slides. Use bullet points, not paragraphs.
  2. Reading from slides. Slides support your talk; they do not replace it.
  3. Skipping the motivation. Explain why anyone should care before diving into the method.
  4. No recorded demo. The demo recording is required and worth 27% of your grade.
  5. Running over time. Practice at least twice. The 7-minute cap is enforced.
  6. Rushing through results. This is the payoff. Show numbers, comparisons, and what you learned.

Submission Checklist

Resources

back