DEV Community

yaksh gandhi
yaksh gandhi

Posted on

The Research Workflow I Wish I Knew in Grad School

I spent three hours every Sunday organizing research I could not find on Monday.

47 browser tabs. Three half-filled notebooks. A Downloads folder that looked like a digital landfill. I was pulling 60-hour weeks, drowning in papers, and still could not locate that one quote I knew I had read somewhere.

My advisor would ask, "What did you think of Chen et al.?" and I would freeze — not because I had not read it, but because I had no idea which of my 200 PDFs it was hiding in.

Then I discovered a workflow that turned my scattered mess into a searchable, listenable, automated knowledge base.

The Breaking Point

It was 2 AM, three weeks before my thesis defense. I needed a specific statistic about renewable energy adoption rates that I had read months ago. I remembered the idea perfectly. I had no idea where I had seen it.

I checked:

  • My browser history (47,000 entries deep)
  • My Downloads folder (a graveyard of paper_final_FINAL.pdf)
  • My Google Drive (organized into folders I had forgotten existed)
  • My physical notebook (flipping through 200 pages of handwriting)
  • My email (searching for "paper" returned 3,847 results)

Two hours later, I gave up. The statistic was gone. I would have to rewrite that section without it.

That night, I realized something: The problem was not my memory. It was my system.

The Workflow: Capture → Process → Consume

Here is the exact system I built.

Step 1: Capture Everything

The first rule: If it is important, it goes in the system immediately. No exceptions.

What How
Research papers Drag PDFs directly, or import from arXiv, Google Scholar, any URL
AI conversations One-click import from ChatGPT, Claude, Gemini, Perplexity
Web articles Highlight anything interesting, snipe it in seconds
Notes & thoughts Voice memos, quick text dumps, screenshots
Social threads Twitter/X discussions, Reddit posts, LinkedIn articles

Step 2: Let Automation Do the Work

Auto-Researcher: Feeds my sources into NotebookLM and generates briefing documents automatically.

Smart Sort: Automatically categorizes sources by topic, date, and relevance.

Generator: Creates study guides, flashcards, and review documents from my sources.

Step 3: Consume However Fits Your Life

Time Activity What I Do
Morning commute (30 min) Listen Podcast episode from yesterday research via private RSS feed
Lunch break (20 min) Review Auto-generated briefing doc — skim for insights
Evening walk (45 min) Listen Another episode, or review flagged sources
Weekend (2 hours) Deep dive Search for patterns across everything I have captured

The Results

Before After
3 hours/week organizing 0 hours (automated)
47 browser tabs open Everything searchable in one place
Lost 30% of AI insights 100% captured permanently
Reading only at desk Learning 5+ hours/week during commute/walks
Could not find quotes Find anything in <10 seconds

Time saved: 5+ hours per week

The Bottom Line

Grad school does not have to be a chaos of lost papers and 2 AM panic searches.

The workflow is simple: Capture everything. Automate the organization. Consume however fits your life.

Your research should work for you. Not the other way around.


This article was written about Kortex, a Chrome extension that enhances NotebookLM with AI chat imports, automations, and private podcast feeds.

Top comments (0)