AI-Sys-Sp22 Machine Learning Systems - University of California, Berkeley vs PostHog

Organizes graduate students into rotating reviewer roles (summarizer, systems expert, applications expert) for weekly research paper evaluations, with written reviews submitted before synchronous discussion sessions. The course implements a mini-program-committee (mini-PC) meeting format where students present papers in different roles across multiple weeks, simulating academic conference review workflows. Reviews are collected asynchronously via email/Slack, then discussed in real-time Monday seminars with instructor feedback.

Unique: Implements rotating reviewer roles (summarizer, systems expert, applications expert) across multiple presentations per student, forcing deep engagement with papers from different analytical angles rather than single-pass reviews. This mirrors academic conference review workflows where reviewers specialize in different aspects.

vs alternatives: More rigorous than traditional lecture-based ML courses because students must defend their analysis in multiple roles; more scalable than one-on-one mentoring but less scalable than MOOCs due to presentation slot constraints.

Schedules industry and academic experts (e.g., Reynold Xin from Databricks) to present on specific ML systems topics during seminar sessions, with speaker slides and discussion topics posted on course website and Slack. Guest speakers are invited to discuss their research or systems work, providing real-world context for the week's assigned papers. Zoom links are posted to Slack for remote attendance, enabling asynchronous participation for students unable to attend live.

Unique: Integrates guest speakers directly into the seminar schedule alongside paper reviews, creating a hybrid learning model where students compare academic research (papers) with practitioner perspectives (speakers) in the same week. This is more integrated than traditional guest lecture series.

vs alternatives: More authentic than recorded expert interviews because speakers can respond to student questions in real-time; more accessible than industry internships because students gain exposure without employment commitment.

Maintains a static course website (hosted on GitHub Pages at ucbrise.github.io/cs294-ai-sys-sp22/) with weekly reading lists, lecture slides, speaker information, and project guidelines. The website is version-controlled via GitHub, allowing instructors to update readings and materials each semester. Students can suggest additional readings via GitHub pull requests, creating a crowdsourced reading list expansion mechanism.

Unique: Uses GitHub as the primary content management system, making the course materials version-controlled and enabling student contributions via pull requests. This treats course content as open-source software rather than proprietary LMS content.

vs alternatives: More transparent and portable than traditional LMS (Canvas, Blackboard) because materials are in plain text and publicly archived; more collaborative than email-based reading distribution because students can propose additions via pull requests.

Uses Slack as the primary communication channel for course announcements, guest speaker Zoom links, office hours coordination, and real-time discussion. Instructors post weekly updates, speaker information, and logistical details to Slack channels; students use Slack threads to ask questions and coordinate. Office hours are arranged via email but discussions may occur in Slack channels.

Unique: Treats Slack as the primary course communication hub rather than a secondary notification channel, centralizing announcements, speaker links, and informal discussion in one platform. This reduces email overhead but increases Slack dependency.

vs alternatives: More responsive than email-based communication because Slack notifications are real-time; less formal than LMS discussion boards, enabling casual peer-to-peer discussion; less persistent than course websites because Slack messages are ephemeral.

Requires students to complete a hands-on project (details not fully specified in provided excerpt) that applies ML systems concepts from the course. Projects are evaluated by instructors and may involve implementation, benchmarking, or systems design. The course encourages mixed teams of AI and systems students to collaborate on projects that bridge both domains.

Unique: Explicitly encourages mixed AI/systems teams, requiring students to bridge academic ML research with systems-level implementation concerns (hardware optimization, distributed training, etc.). This is more integrated than separate AI and systems projects.

vs alternatives: More practical than paper-only seminars because students must implement and benchmark systems; more flexible than structured labs because students design their own projects; less guided than bootcamp-style courses because project scope is student-defined.

Provides Zoom links (posted on Slack) for students unable to attend Monday seminars in person, enabling synchronous remote participation in paper discussions and guest speaker sessions. Zoom attendance is optional but encouraged; no asynchronous recording or replay is documented. Remote students can participate in real-time Q&A and discussion.

Unique: Treats Zoom as a synchronous participation channel rather than a recording/replay mechanism, maintaining the seminar's real-time discussion culture while accommodating remote students. This is more inclusive than in-person-only but less accessible than recorded lectures.

vs alternatives: More engaging than asynchronous video because students can ask real-time questions; less accessible than recorded lectures because students must attend live; simpler to manage than hybrid breakout rooms because all participants are in one Zoom meeting.

Instructors offer office hours by email arrangement (no fixed schedule documented), allowing students to request one-on-one meetings to discuss papers, projects, or course content. Students email instructors to schedule meetings; office hours may occur in-person or via Zoom depending on student preference and instructor availability.

Unique: Uses email as the primary office hours scheduling mechanism rather than a calendar system (Calendly, Google Calendar, etc.), creating a more personal but less scalable approach. This reflects the seminar's intimate, low-tech culture.

vs alternatives: More flexible than fixed office hours because students can request meetings at any time; less scalable than calendar-based scheduling because coordination is manual; more personal than automated scheduling because instructors can customize meeting format and duration.

PostHog/posthog | DeepWiki Loading... Index your code with Devin DeepWiki DeepWiki PostHog/posthog Index your code with Devin Edit Wiki Share Loading... Last indexed: 28 May 2026 ( 4a5e38 ) Overview Monorepo Structure and Build System Frontend Workspace and Product Packages Python Dependencies and Configuration CI/CD Pipeline Schema and Type System Cross-Language Schema Synchronization Query Schema Definitions Database Migrations Data Storage and Ingestion ClickHouse Architecture Kafka to ClickHouse Pipeline PostgreSQL and Database Pools Query Log Archive System Event Ingestion Pipeline (Node.js) Backend Services Django Middleware System Feature Flags Service (Rust) API Layer and Authentication Rust Microservices LLM Gateway Service Agentic Provisioning and OAuth Max AI Assistant Architecture and Agent Modes Query Execution and Streaming Frontend Integration MCP Server Tasks (AI Coding Agent) Feature Flags System Feature Flag Management API Flag Evaluation and Dependencies Frontend Interface Product Features Logs Viewer Session Recordings Insights and Analytics Surveys and Scheduled Changes Experiments (A/B Testing) Web Analytics Error Tracking LLM Analytics Frontend Architecture Kea State Management Product Module System Build System and Tooling Testing and Quality Test Infrastructure Backend and Rust Tests Frontend and E2E Tests Data Platform and Workf

Monorepo Structure and Build System | PostHog/posthog | DeepWiki Loading... Index your code with Devin DeepWiki DeepWiki PostHog/posthog Index your code with Devin Edit Wiki Share Loading... Last indexed: 28 May 2026 ( 4a5e38 ) Overview Monorepo Structure and Build System Frontend Workspace and Product Packages Python Dependencies and Configuration CI/CD Pipeline Schema and Type System Cross-Language Schema Synchronization Query Schema Definitions Database Migrations Data Storage and Ingestion ClickHouse Architecture Kafka to ClickHouse Pipeline PostgreSQL and Database Pools Query Log Archive System Event Ingestion Pipeline (Node.js) Backend Services Django Middleware System Feature Flags Service (Rust) API Layer and Authentication Rust Microservices LLM Gateway Service Agentic Provisioning and OAuth Max AI Assistant Architecture and Agent Modes Query Execution and Streaming Frontend Integration MCP Server Tasks (AI Coding Agent) Feature Flags System Feature Flag Management API Flag Evaluation and Dependencies Frontend Interface Product Features Logs Viewer Session Recordings Insights and Analytics Surveys and Scheduled Changes Experiments (A/B Testing) Web Analytics Error Tracking LLM Analytics Frontend Architecture Kea State Management Product Module System Build System and Tooling Testing and Quality Test Infrastructure Backend and Rust Tests Frontend a

Schema and Type System | PostHog/posthog | DeepWiki Loading... Index your code with Devin DeepWiki DeepWiki PostHog/posthog Index your code with Devin Edit Wiki Share Loading... Last indexed: 28 May 2026 ( 4a5e38 ) Overview Monorepo Structure and Build System Frontend Workspace and Product Packages Python Dependencies and Configuration CI/CD Pipeline Schema and Type System Cross-Language Schema Synchronization Query Schema Definitions Database Migrations Data Storage and Ingestion ClickHouse Architecture Kafka to ClickHouse Pipeline PostgreSQL and Database Pools Query Log Archive System Event Ingestion Pipeline (Node.js) Backend Services Django Middleware System Feature Flags Service (Rust) API Layer and Authentication Rust Microservices LLM Gateway Service Agentic Provisioning and OAuth Max AI Assistant Architecture and Agent Modes Query Execution and Streaming Frontend Integration MCP Server Tasks (AI Coding Agent) Feature Flags System Feature Flag Management API Flag Evaluation and Dependencies Frontend Interface Product Features Logs Viewer Session Recordings Insights and Analytics Surveys and Scheduled Changes Experiments (A/B Testing) Web Analytics Error Tracking LLM Analytics Frontend Architecture Kea State Management Product Module System Build System and Tooling Testing and Quality Test Infrastructure Backend and Rust Tests Frontend and E2E Tests

PostHog/posthog | DeepWiki Loading... Index your code with Devin DeepWiki DeepWiki PostHog/posthog Index your code with Devin Edit Wiki Share Loading... Last indexed: 28 May 2026 ( 4a5e38 ) Overview Monorepo Structure and Build System Frontend Workspace and Product Packages Python Dependencies and Configuration CI/CD Pipeline Schema and Type System Cross-Language Schema Synchronization Query Schema Definitions Database Migrations Data Storage and Ingestion ClickHouse Architecture Kafka to ClickHouse Pipeline PostgreSQL and Database Pools Query Log Archive System Event Ingestion Pipeline (Node.js) Backend Services Django Middleware System Feature Flags Service (Rust) API Layer and Authentication Rust Microservices LLM Gateway Service Agentic Provisioning and OAuth Max AI Assistant Architecture and Agent Modes Query Execution and Streaming Frontend Integration MCP Server Tasks (AI Coding Agent) Feature Flags System Feature Flag Management API Flag Evaluation and Dependencies Frontend Interface Product Features Logs Viewer Session Recordings Insights and Analytics Surveys and Scheduled Ch