Build your portfolio.
Real work. Real companies.

A live dashboard powered by your AI agents, built to surface trends, benchmarks, and content suggestions that Wayfair’s category team can actually use.

‍

Project Breakdown:

1. Build a Moodboard Agent: Create a simple AI tool that transforms style prompts into curated visual outputs.

2. Discover Market Trends: Build agents that collect and analyze product data, blogs, and social media to uncover design trends.

3. Monitor Competitors: Develop AI agents to track competitor launches, pricing updates, and marketing campaigns.

4. Generate AI Insights: Turn raw trend signals into summaries, draft blog ideas, and campaign captions.

5. Build the Dashboard: Combine datasets and insights into a live, auto-updating Google Sheets tool for managers.

‍

Tools Used:

n8n - a workflow automation software tool

Google Gemini

Mistral AI

Wayfair is one of the world’s largest online destinations for home furnishings and is known for its data-driven approach. Behind the 30M+ products you see online are teams using AI to move faster and smarter. This is your backstage pass to see — and build — how that happens.

Provide Breaking Games with clear, actionable strategies based on the data analysis, such as optimizing ad targeting, improving website flow, and/or tailoring promotional efforts to enhance sales.

• E-Commerce Sales Funnel Report: A summary of key metrics and marketing channel performance.
• Visitor Behavior Dashboard: Analysis of traffic sources and engagement with actionable improvement suggestions.
• Paid Campaign Performance Report: Identification of top-performing ads with ROI-based budget reallocation strategies.
• Conversion Funnel Analysis: Detailed breakdown of drop-off points and product performance insights.
• Cross-Channel Attribution Report: Evaluation of how various marketing efforts directly drive Shopify sales.
• Visual Customer Journey Map: Diagnosis of funnel bottlenecks from initial ad exposure to final purchase.
• Customer Segmentation Plan: Data-driven marketing strategies for high-value upselling and retargeting.
• Final Strategic Action Plan: A professional presentation synthesizing all insights into a comprehensive growth roadmap with defined KPIs.

Breaking Games is the publishing arm of AdMagic—the powerhouse behind hits like Cards Against Humanity and Exploding Kittens. But they’re more than just a traditional publisher. They help invent, test, launch, and distribute both their own games and others’, working across the entire tabletop ecosystem. This externship offers a rare peek behind the scenes of how digital marketing powers the business side of play.

During this Externship, externs will:

• Ingest and parse scanned documents
• Extract and structure key data fields from multiple document types
• Implement intelligent search using RAG pipelines and metadata filters
• Build and test modular routing and classification logic
• (Optional) Deploy a lightweight user interface for search and querying

‍

Technologies used:

• Python (scripting, data processing)
• PyMuPDF, pdfplumber (PDF parsing)
• OpenCV, PIL (image preprocessing)
• OCR Engines: Tesseract, PaddleOCR, EasyOCR
• Vector Search: LlamaIndex, FAISS, Chroma
• LLMs: Gemini API or open-source models (e.g., Mistral, Phi-2)
• UI Tools: Streamlit, Gradio (optional)

‍

Final project outputs:

• Python pipeline to process document files
• Structured JSON output with field-level data and coordinates
• Functional RAG pipeline (Gemini or open-source)
• Searchable UI using Streamlit or Gradio
• Final report, system documentation, and demo

‍

Why should you care? Here’s what you’ll get out of this Externship:

• Develop a portfolio-worthy AI/ML project
• Gain hands-on experience with cutting-edge tools and models
• Learn to build robust, production-level systems
• Improve data parsing, extraction, and visualization skills
• Tackle a real-world, high-impact business use case

Pfizer is one of the world's largest publicly traded pharmaceutical companies, developing and manufacturing medicines and vaccines that reach hundreds of millions of people globally. They partnered with BioNTech to create the first FDA-authorized COVID-19 vaccine. In 2024, Pfizer reported $63 billion in revenue and products reaching 414 million patients worldwide—massive scale. Their clinical supply operations ensure that the right drugs get to the right patients at the right time—a process that depends on flawless documentation across shipping, logistics, and regulatory compliance.

• Pick your focus area and go deep: Choose one focus area within energy and dig into it until you understand it better than almost anyone: the problem, who it affects, and why it matters now. That deep dive makes you the go-to person on your space and sets up everything you build next.
• Map the technologies solving it: Find the technologies tackling your focus area, label each by how mature it really is, and flag the most promising ones to carry forward.
• Source the startups in the space: Research startups chasing those technologies and organize them into one clean, comparable table.
• Build the market landscape: Turn your list into a landscape map, then write the competitive readout that names who is ahead and where the open space is.
• Size the opportunity: Estimate the market size from several sources (the whole market, the reachable slice, and the realistic share), and stay honest when the real number is smaller than the headline.
• Make your investment recommendation: Pick the most investable company from your own map, build the evidence, weigh the risks, and make the case for why EIC should back it, in a VC deal-memo format.

Energy Innovation Capital (EIC) is a venture capital firm investing in companies developing the next generation of technology to ensure abundant, clean, and accessible energy. With offices in San Francisco, Houston, Boston, and New York, the EIC team is made up of multiple-time CEOs and experienced technology investors who have collectively led over 150 investments across the energy tech landscape. To date, EIC has backed more than 30 early-stage companies through three funds, supporting founders focused on innovation in energy transition, digital technologies, and sustainability.

• Get Familiar with the Data: Learn the fundamentals of data analytics and its applications in healthcare. Use AI tools to refine research questions and explore public datasets.
• Data Collection & Pre-processing: Gather infection-related data from public sources, clean and structure it for analysis, and address missing values.
• Exploratory Data Analysis (EDA): Perform statistical and correlational analyses to identify patterns and trends in infection rates. Create visual representations of findings.
• Presentation Development: Develop structured presentations using AI-generated reports and visualization techniques. Use Gamma.app and Google Slides for final presentations.
• Interactive Dashboard Creation: Build interactive dashboards with Power BI and Claude to showcase trends and facilitate data-driven decision-making.

TruBridge has over four decades of experience in connecting providers, patients and communities with innovative solutions that create real value by supporting both the financial and clinical side of healthcare delivery.

‍

We are a trusted partner to healthcare organizations with a broad range of technology offerings that address the unique needs and challenges of diverse communities, promoting equitable access to quality care and fostering positive outcomes.

‍

Our industry leading HFMA Peer Reviewed® suite of revenue cycle management (RCM) offerings combine unparalleled visibility and transparency to enhance productivity and support the financial health of healthcare organizations across all care settings.

‍

Our solutions champion end-to-end, data-driven patient journeys that support value-based care, improve outcomes, and increase patient satisfaction. We support efficient patient care with electronic health record (EHR) product offerings that successfully integrate data between care settings.

‍

Above all, we believe in the power of community and encourage collaboration, connection, and empowerment with our customers. We clear the way for care.

• Design and Run a Real Survey: Deploy it online and face to face, and collect real responses from at least 20 people.
• Analyze What You Actually Collected: Turn raw survey and interview data into real, defensible patterns.
• Benchmark the Competition: Map Generous Brands against the biggest names in the category.
• Build a Consulting-Style Recommendation: Use the Pyramid Principle to argue your case and end with a real next step.

Generous Brands, formed in 2024, owns a portfolio of premium refrigerated beverage brands built around Bolthouse Farms, Evolution Fresh, and SAMBAZON, with Health-Ade joining as its newest addition, together nearing a billion dollars in retail sales. Health-Ade itself started as a kombucha sold at a farmers market in 2012 before growing into one of the top-selling functional beverage brands in the country.

Develop a comprehensive creative strategy framework for Beats by Dre video advertising, then demonstrate its application through an original creative concept.

‍

• Catalog and analyze 20-30 video ads across TikTok, Instagram, and YouTube into a structured, insight-ready repository.
• Design and execute a consumer survey to capture real reactions to advertising content, combined with social listening insights.
• Build a Creative Strategy Framework including platform best practices, brand creative standards, effectiveness rubric, competitive analysis, and strategic recommendations.
• Develop an original creative concept with a creative brief, framework compliance evaluation, and a 5-minute strategic presentation.

‍

By the end of this Externship, externs will have created:

• Ad Repository Database (20-30 tagged ads)
• Consumer Feedback Report
• Creative Strategy Framework Package
• Creative Concept Presentation

Beats by Dre, founded in 2006, is a leading audio brand known for its premium headphones, earphones, and speakers. As a company focused on innovation and customer satisfaction, Beats by Dre uses data-driven insights to ensure its products stay ahead of market trends and meet the needs of its global audience.

Analyze a real multi-file dataset spanning 300+ juvenile justice facilities to develop research questions, conduct statistical analyses in Python, and deliver a professional findings package — including a written brief, data visualizations, presentation deck, and recorded presentation — that goes directly into CIYJ’s advocacy work with state agencies and legislators.

Deliver a professional FP&A consulting package and company valuation for Attronica:

1. Organize and analyze historical financial statements: Review Attronica’s audited financials, restructure data for clarity, and identify key trends across revenue, margins, and operating expenses.
2. Build a financial forecasting model: Develop a forward-looking financial model in Excel incorporating revenue drivers, cost assumptions, and goal-seek analysis to project multiple growth scenarios.
3. Conduct operational and competitive analysis: Evaluate Attronica’s market position within the IT hardware and services landscape, benchmarking against industry peers and identifying strategic advantages.
4. Produce a comprehensive private company valuation: Apply DCF, comparable company, and precedent transaction methodologies to arrive at a defensible valuation range supporting strategic business decisions.
5. Deliver a final strategy presentation to the CEO: Synthesize all findings into a polished slide deck presenting the FP&A assessment, valuation summary, and strategic recommendations for Attronica’s next chapter.

Design and prototype a feature aimed at boosting BeReal's monthly active users metric (MAUs). Demo and present to BeReal leadership at the end of the externship.

‍

Project deliverables include:

1) A visual representation of BeReal’s user behavior pattern highlighting common usage habits, motivations, and frustrations of BeReal users

2) A competitor analysis matrix highlighting engagement-driving features and gaps

3) Prototype Flow Map (photo or text storyboard) and Prototype Link (+ README with click path)

4) Focus group test summary report

5) Final feature proposal deck (5–7 slides) presenting the feature, supporting evidence, and rollout plan

Founded in France, BeReal is a fast-growing social app with 40M users worldwide. They’re changing the way people connect—by encouraging users to share a real, unfiltered glimpse of their lives every day. In a landscape dominated by filters and algorithms, BeReal is fighting to let us all be “real” again.. BeReal is betting big on genuine connection—and wants your help in crafting features that stay true to that mission while growing engagement and reach.

Deliver a working analytics database and a clear, data-backed recommendation that tells Breaking Games where to spend and where to pull back.

• Load and explore the raw data: Get six separate data files into a database and write your first queries — feeling firsthand why messy, disconnected data is so hard to work with.
• Design a database that connects the data: Structure proper tables, link them with keys, and join them to follow one customer’s journey from first ad click to purchase.
• Find out what’s actually selling: Rank the full product catalog by what drives revenue, build summary tables, and turn the numbers into a clear restock recommendation.
• See which marketing is working: Clean hundreds of messy traffic sources into tidy channels, then calculate which ads and channels actually pay off.
• Investigate why carts get abandoned: Dig into thousands of abandoned carts, uncover which “shoppers” are bots skewing the numbers, and calculate the real abandonment rate.
• Build a dashboard and make your recommendation: Pull everything into a clean, shareable HTML dashboard and make your case to the team.

Breaking Games is the publishing arm of AdMagic—the powerhouse behind hits like Cards Against Humanity and Exploding Kittens. But they’re more than just a traditional publisher. They help invent, test, launch, and distribute both their own games and others’, working across the entire tabletop ecosystem. This externship offers a rare peek behind the scenes of how digital marketing powers the business side of play.

During the course of the externship, externs will:

• Analyze how case management workflows influence patient experience, continuity of care, and risk across inpatient, emergency, and post-discharge contexts.
• Identify process gaps, blind spots, and unaccounted risks by applying root cause analysis to real care coordination scenarios.
• Map real-world case management workflows to reveal hidden work, handoffs, delays, and intervention points.
• Quantify coordination work by defining operational metrics and prototyping a dashboard that communicates workload, risk mitigation, and value.
  
  

By the end of the externship, externs will have created a:

• Stakeholder map and scope-of-practice summary clarifying the case manager’s role in patient-centered care coordination.
• Workflow risk and root-cause analysis table highlighting gaps, risks, and invisible coordination work.
• Process maps and visual flowcharts showing how case management work actually unfolds across follow-up, default management, and escalation.
• Work-quantification dashboard prototype that translates coordination activities into measurable, communicable operational value.

‍

Tools being used: Claude, Perplexity, Gamma

The Institute of Mental Health (IMH) is Singapore’s national acute psychiatric hospital, with over 1,950 beds and a mission to deliver holistic, integrative mental healthcare. It serves children, adolescents, adults, and the elderly through multidisciplinary teams that include psychologists, nurses, occupational therapists, and social workers. IMH combines evidence-based care with operational excellence to support mental health at scale.

Create a professional slide deck and present your investment thesis to Mangusta's Partners:

‍

1. Source early stage startups you want to pitch: Search for and identify early-stage startups through diverse discovery methods to uncover under-the-radar companies for further evaluation.
2. Conduct comprehensive due diligence: Gather and analyze company data, including fundraising history, team composition, press coverage, and financial metrics, to build reports for investment assessment.
3. Generate and refine a targeted list of key competitors: Conduct an in-depth analysis of competitive positioning and differentiation to evaluate a startup’s potential within its market.
4. Analyze B2C traction data: Interpret key metrics such as user engagement and app downloads. Leverage and develop AI-driven strategies to help synthesize customer sentiment data to identify product strengths and areas for improvement.
5. Complete comprehensive startup analysis on core areas: This includes Market Opportunity, Technology & Core IP, Competitive Landscape, Go-to-Market Strategy, Team Assessment, and Risk Analysis, translating due diligence findings into actionable insights.
6. Produce a strategic synergy report: This report should align the startup’s potential with the VC fund’s objectives, articulating collaborative value and strategic fit.
7. Design and deliver a concise 3-slide investment memo: Deliver a polished, data-driven presentation in Google Slides to communicate key findings to potential investors as you would if you were a VC associate on the team.

Mangusta Capital is a venture capital and growth equity firm with origins in Italy, founded by SoftBank Vision Fund’s founder and backed by the family behind Ray-Ban/Oakley. The mission is to shape the world for future generations by supporting technology innovations that move humanity forward.

They invest in audacious founders building mission-critical capabilities across AI and software. As they reach the tipping point in the advent of AGI, we see the inevitability of unlocking massive productivity gains in complex workflows across all industries.

They invest in the next generation of great consumer brands and technologies enabling delightful consumer experiences. In the past, consumers trusted brands; today, consumers trust influencers. Mangusta searches for the trendsetters changing the way we eat, play, and live.

Produce a complete, self-sourced children's environmental health findings package:

1. Evaluated Dataset: a real-world public health dataset you sourced, evaluated for fitness, and cleaned for analysis
2. Defended Research Question: a focused, testable question about environmental exposures and child health outcomes in a community of your choice
3. Statistical Analysis: correlation and regression analysis run in Python, with method-selection rationale and honest interpretation
4. Stress-Tested Finding: a revised interpretation that survives a structured pitfalls checklist covering confounders, ecological fallacy, and overclaiming
5. Practitioner-Facing Presentation: a written brief, supporting chart, slide deck, and recorded presentation built for a non-technical public health audience