Coming soonSoftware & AIMastery

Advanced Postgres — Indexes, Plans, and Tuning

Diagnose and fix the slow queries that cost MENA companies $thousands in unnecessary RDS spend. Real explain plans, real partial indexes, real performance gains.

About this course

When your Postgres database is slow, the answer is rarely 'add more RAM.' It's usually a missing index, a bad query plan, or a misunderstood VACUUM. We diagnose real cases (sanitized from production at three MENA startups), measure with auto_explain, fix with proper indexes (B-tree, GIN, partial, expression), and explain why each fix worked. By the end you can read any explain plan and recommend a tuning that survives a code review.

What you'll cover

1
How Postgres executes a query
Parser, planner, executor. The shape of an EXPLAIN plan.
2
Indexes — B-tree, GIN, partial, expression
Which index for which query. The cost of an index nobody mentions.
3
EXPLAIN ANALYZE deeply
Cost vs actual. Buffers. The 5 patterns that always mean trouble.
4
Statistics, ANALYZE, and the planner's mind
How Postgres guesses row counts. When it guesses wrong. Forcing a better plan.
5
VACUUM, autovacuum, and bloat
Why your table is 50GB but holds 5GB of data. Tuning autovacuum.
6
Connection pooling and pgbouncer
Why max_connections=200 isn't the answer. Transaction-mode pooling.
7
Real cases — diagnose, fix, measure
5 sanitized production slow queries. Each one is a 30-minute exercise.

Who it's for

Senior backend engineers, DBAs, SREs, and tech leads owning database performance.

Prerequisites

Comfortable with SQL — JOINs, GROUP BY, subqueries. You've worked with Postgres in production.

Skills you'll build

PostgreSQL
EXPLAIN
indexes
query tuning
VACUUM
pgbouncer
DBA
performance

Who we're looking for

Open call · Apply to teach

Required skills

PostgreSQL
EXPLAIN
indexes
query tuning
VACUUM
pgbouncer
DBA
performance

Experience

5+ years senior practitioner

Languages

English or Arabic (both a plus)

Time commitment

8 sessions × 2 hours over 8 weeks

Compensation

80% of seat revenue (Tahout takes 20%)

If your CV matches, apply to teach. We use AI to rank applicants by fit, then admin reviews and approves the right instructor(s).