I mapped roles and responsibilities across an Amazon Fulfillment Center, documenting each role's main duties, interactions, tools, and why it mattered.

I cleaned and structured employee feedback from Glassdoor and YouTube, producing an analysis-ready dataset and examples of strong employee voice.

I cleaned and prepared employee feedback text for analysis, applying tokenization, stopword removal, stemming, and noted issues with negation handling.

I identified themes in employee feedback, charted sentiment by theme, and traced root causes for the top issue using 5 Whys and visual trees.

I mapped BeReal’s end-to-end user journey and produced a journey map file showing stages, behaviors, and pain points.

I compiled a side-by-side product feature comparison between BeReal and a selected competitor, scoring themes and summarizing insights in a shared comparison table.

I designed and prototyped a Caption Prompts feature that added optional, lightweight caption suggestions to encourage expressive, authentic posts.

I designed and ran a focus group testing a caption feature, drafted the discussion guide and survey, facilitated sessions, and synthesized participant feedback into recommendations.

I identified research questions linking lifestyle and cancer, searched public sources, and selected datasets to support analysis.

I collected a public cancer risk factors CSV, verified its structure, and prepared it for analysis.

I performed exploratory data analysis and statistical tests on a health dataset, built visualizations, and reported model metrics and significance tests.

I drafted a 10–15 minute presentation outline summarizing analyses of smoking, air pollution, and cancer risk, with visuals and policy recommendations.

I picked a semiconductor stock and narrowed my selection to Siltronic AG after screening and comparing candidates.

I summarized press releases, 10-K highlights, and financial statements to assess the company’s recent performance and risks.

I completed a competitor mapping and timeline for the selected company, listing direct and adjacent wafer producers and recent company milestones.

I built a financial model and populated historicals and forecasts for my selected company, including income statement, balance sheet, segment detail, and quarterly forecasts in Excel/Sheets.

I summarized how large language models work, from tokenization to training and fine-tuning, in the Research and Explain How LLMs Work task.

I prepared and cleaned pharmaceutical documents using Python in Google Colab, practicing data structures, control flow, functions, Pandas, JSON handling, text cleaning, and image preprocessing.

I extracted and structured text and keywords from multi-page SDF PDFs using Python tools and shared the working notebook and report.

I compared Tesseract, PaddleOCR, and EasyOCR on scanned pharmaceutical supplier pages and reported which engine handled layout and codes best.

I built and tested a RAG pipeline for document retrieval, producing embeddings, chunking, and a hybrid retriever and validating outputs with test queries.

I ran side-by-side evaluations of Gemini and open-source models on SDF document queries and compared their answers and timing.

I built a segmentation and metadata pipeline to split bundled PDFs into document-level chunks and tag each chunk for routing.

I built a Gradio-based chatbot UI and connected it to a local RAG pipeline that accepted uploaded documents and returned context-aware answers.