Archaeologist·Blog Dispatch — Dropbox Tech Blog

Vol. I · Field Notes

DropboxDropbox Tech Blog

Dropbox Tech Blog is the engineering voice of a company that started as a file-sync service and now builds AI-powered search and knowledge tools. The blog covers storage optimization, AI infrastructure, and developer productivity. It's best for engineers curious about how a large-scale SaaS company evolves its core systems and experiments with LLMs.

9 May 2026·10 posts·5 clusters·10 authors

⁂

Reading Posture

From the Field

“Dropbox engineering scales storage and AI with pragmatic deep dives.”

Verdict:Reach for it

Reach for it when

Read this when you want to learn how a mature infrastructure company optimizes storage, adopts LLMs for search, and balances developer velocity with reliability.

Look elsewhere when

Skip it if you need beginner tutorials, frontend engineering, or frequent incident post-mortems.

In context

Compared to Uber Engineering, this one is more focused on storage internals and AI integration rather than microservices at scale.

Complexity●●●Heavy

Read time~120 minutes

Language

Blog

Runtime

web

Dependencies

0total

What this is

As told for the tourist

Dropbox Tech Blog is the engineering voice of a company that started as a file-sync service and now builds AI-powered search and knowledge tools. The blog covers storage optimization, AI infrastructure, and developer productivity. It's best for engineers curious about how a large-scale SaaS company evolves its core systems and experiments with LLMs.

Start Here

A recommended reading path through the code

Start Here

A recommended reading path through the code

01
Start here because it's a classic Dropbox strength—deep storage internals with clear trade-offs.
02
Shows their modern AI approach with prompt optimization and LLM pipelines.
03
Connects infrastructure to AI, explaining real-time feature serving.
04
Covers observability at scale, a key operational concern.
05
Reveals developer velocity strategies and internal tooling philosophy.
06
Practical optimization for LLM serving, relevant for ML engineers.
07
Advanced architectural view of their AI assistant, tying together storage, search, and context.

What's inside

5 sections of the codebase

Posting History

Activity over time

Posting Activity10 posts · 2025-11 → 2026-04

2025

3 posts

2026

7 posts

Less

More

The Archive

Every post, searchable and filtered

All Posts10 of 10

Improving storage efficiency in Magic Pocket, our immutable blob store

Improves storage efficiency in Magic Pocket by turning compaction into a layered, adaptive pipeline with enhanced monitoring.

Storage & Infrastructure#deep-dive#infraFacundo Agriel

Reducing our monorepo size to improve developer velocity

Reduces monorepo size to improve developer velocity without increasing friction.

Storage & Infrastructure#dx#scalingFacundo Agriel,Ishan Mishra

How we optimized Dash's relevance judge with DSPy

Optimizes Dash's relevance judge using DSPy to automate prompt engineering and improve performance and cost.

AI Search & Ranking#ml-infra#performanceFacundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon

Using LLMs to amplify human labeling and improve Dash search relevance

Uses LLMs to amplify human labeling for training Dash's search ranking models.

AI Search & Ranking#ml-infra#aiFacundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon,Dmitriy Meyerzon

How low-bit inference enables efficient AI

Explains how low-bit inference enables efficient AI for products like Dropbox Dash.

AI Efficiency & Productivity#performance#ml-infraFacundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon,Dmitriy Meyerzon,Hicham Badri,Appu Shaji

Insights from our executive roundtable on AI and engineering productivity

Shares insights from an executive roundtable on AI and engineering productivity, including adoption of AI coding tools.

AI Efficiency & Productivity#cultureFacundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon,Dmitriy Meyerzon,Hicham Badri,Appu Shaji,Craig Wilhite

Engineering VP Josh Clemm on how we use knowledge graphs, MCP, and DSPy in Dash

Engineering VP discusses how knowledge graphs, MCP, and DSPy are used in Dash.

Dash Architecture & Context#deep-dive#architectureFacundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon,Dmitriy Meyerzon,Hicham Badri,Appu Shaji,Craig Wilhite,Josh Clemm

Inside the feature store powering real-time AI in Dropbox Dash

Describes the feature store powering real-time AI for ranking and retrieval in Dropbox Dash.

AI Search & Ranking#infra#ml-infraFacundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon,Dmitriy Meyerzon,Hicham Badri,Appu Shaji,Craig Wilhite,Josh Clemm,Jason Shang,Artem Nabirkin

Building the future: highlights from Dropbox’s 2025 summer intern class

Highlights the 2025 summer intern class and Dropbox's intern program focused on growth and innovation.

Intern Program & Culture#cultureFacundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon,Dmitriy Meyerzon,Hicham Badri,Appu Shaji,Craig Wilhite,Josh Clemm,Jason Shang,Artem Nabirkin,Dropbox Team,Ameya Bhatawdekar

How Dash uses context engineering for smarter AI

Explains how Dash uses context engineering to help AI models focus on relevant information.

Dash Architecture & Context#deep-dive#aiFacundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon,Dmitriy Meyerzon,Hicham Badri,Appu Shaji,Craig Wilhite,Josh Clemm,Jason Shang,Artem Nabirkin,Dropbox Team,Ameya Bhatawdekar,Sean-Michael Lewis

Export & Share

Take the field notes with you