Skip to main content

Mapping Curriculum Strengths Across Tracks

What This Concept Is​

Across ten semesters you built in many areas -- algorithms, systems, databases, distributed systems, security, ops, architecture. Almost no one ends this degree equally strong across all of them. The concept here is an honest audit: a grid showing where your actual strength pulled, not where you wish it had.

The audit has two axes:

  • Track -- the six common specialization tracks (backend, distributed systems, platform, infra, data, security).
  • Evidence -- for each, the concrete artifacts (capstone subsystems, deep-dives, incidents, projects, writeups) that honestly support claiming strength.

If a cell has no artifact, the claim of strength does not yet exist. This is not a judgment; it is a map.

The audit is deliberately retrospective. It does not ask what you want to be strong in. It asks what you actually produced in the last ten semesters, and where that production clusters. The answer often surprises people -- the track you are strongest in is rarely the track you identified with going in. That mismatch is the single most valuable output of this concept, because it anchors everything that follows (concepts 11 and 12) in evidence rather than identity.

There is also a calibration purpose. A degree where every semester was "fine" but none were load-bearing produces a flat audit -- no strong rows, no weak rows. That is diagnostic: it usually points at work that was completed but never consolidated into something pinnable. The remedy is not to rewrite the audit; it is to consolidate before auditing.

A related failure mode is the sloped audit -- every row rated 3, which reads as modest but is usually an evasion. Honest audits produce a jagged shape: one or two strong rows, one or two weak rows, a couple in the middle. A flat 3-across-the-board audit is almost always unexamined; force the top and bottom rows to move by at least one notch on a second pass, using only artifacts as justification.

Why It Matters Here (In the Capstone)​

The specialization decision (concept 12) fails when made from aspiration. Concepts 10 and 11 exist to replace that aspiration with evidence. A pick made from a real map is defensible in interviews, survives first contact with the job, and gives the 12-month plan (concept 12) something to aim at.

This is where the whole degree gets audited honestly. The outputs of Semesters 2, 4/5, 6, 7/8, and 9 are the source data. A thin grid means the degree produced output the curriculum expected, but that output was not internalized into a direction. The capstone is the last and best chance to close that gap.

The audit also interacts with the portfolio (cluster 2). The tracks the grid ranks highest should be the tracks your pinned repos represent. If the grid says "distributed" is strongest but your pins are half web frontends, one of the two is lying.

Concrete Example: Strength Map Grid​

One row per track. One column per evidence type. Fill in concrete artifact names -- do not leave cells vague.

TrackCourse / module artifactsCapstone evidenceOutside-of-degree workSelf-rating (1-5)
BackendS4 persistent store, S5 API projectcapstone inventory service--4
DistributedS6 Raft drill, S8 consensus ADRcapstone dedup across feedssmall toy gossip protocol3
Platform (dev tooling)S7 CI/CD modulecapstone CI pipeline--2
Infra (cloud, networking)S9 IaC projectcapstone terraform + deploysmall AWS side project3
Data (pipelines, analytics)S3 stats, S5 storageminor feed-ingest work--2
SecurityS4 authn modulethreat model for capstone--2

The self-rating is calibrated: 1 = I know the vocabulary; 3 = I have shipped one working thing; 5 = I could teach it. No one fills this in and scores 5 everywhere.

Concrete Example: Calibration Anchors​

The scale is worthless without concrete anchors. Use these:

  • 1 -- Vocabulary. You could follow a conversation; you could not lead one. Example: you have heard of Paxos but never implemented a quorum.
  • 2 -- Tutorial. You have followed a tutorial end-to-end and modified the output slightly. Example: you have run one node of Kafka locally.
  • 3 -- Shipped once. You have built one working version of the thing, with failure handling, and operated it for at least a week. Example: you ran a three-node distributed system and debugged a partition.
  • 4 -- Shipped under pressure. You have built the thing, operated it through at least one real incident, and written about it. Example: you wrote the runbook your team used during an outage.
  • 5 -- Could teach. You have mentored another engineer through building the thing and produced a writeup others have used. Example: you have delivered a talk or written a post that someone cited.

Re-rate every cell against these anchors. Most initial ratings drop by one level on honest recalibration; that is the correct direction.

Common Confusion / Misconceptions​

Rating-by-interest. If you are fascinated by distributed systems but have built one toy Raft implementation, you are a 2 in distributed systems with a pull toward it. That is different from a 4. Interest is a track input, not a rating input.

Confusing familiarity with strength. Reading about Kubernetes is not the same as running it in production. The grid must reflect the second, not the first. Passive knowledge does not rate higher than 2 regardless of how much you have read.

Skipping the self-rating because it feels arbitrary. The rating is imprecise on purpose. Its job is to force a single-digit decision per cell, which surfaces the distribution, which is what the next concepts need.

Padding cells with tangential work. A one-week hack is not evidence for the track; it is evidence that you can hack. The grid measures sustained output, which is why the rating anchors emphasize operation, not just shipping.

How To Use It (In Your Capstone)​

Fill the grid in this order:

  1. List the tracks (concept 11 will define them more precisely).
  2. For each track, write down every concrete artifact from the degree -- module name plus what you actually produced.
  3. Include non-curriculum work only if it is defensible (shipped, observable).
  4. Rate 1-5, calibrated against the anchors above.
  5. Mark the top two and bottom two tracks. Those are the signal.
  6. Share the grid with a peer from the program; ask them to challenge any rating where the artifact-to-rating ratio looks off.
  7. Commit the final grid to your portfolio repo under library/raw/strength-grid.md; re-run it every 12 months.

Check Yourself​

  1. Does every non-zero rating have at least one concrete artifact behind it?
  2. Which two tracks hold the most artifacts? Which two hold the fewest?
  3. Do the two you rate highest also correspond to the work you most enjoyed? If not, where is the gap?
  4. Did any rating drop when you applied the calibration anchors? That is the useful kind of correction.
  5. Has a peer challenged any cell and pushed back successfully? If not, the audit may still be optimistic.
  6. Does the grid's top row match the shape of your pinned portfolio from cluster 2? If not, one of them is off.
  7. For each cell at rating 3 or higher, is there a specific artifact name (not a topic area) you can cite within five seconds?
  8. Has the grid been re-audited within the last 12 months? If older, rerun now; the data has shifted.

Mini Drill or Application (Capstone-scoped)​

Three drills:

  1. Populate from memory. Build the grid for yourself now. Use the six rows above. Do not edit them to flatter yourself. Leave cells empty where there is no real artifact.
  2. Anchor recalibration. Re-rate every cell using the 1-5 anchors in this concept. Track which cells dropped; those are the ones where familiarity was being confused with strength.
  3. Peer challenge. Share the grid with one peer. Ask them to challenge any rating where the artifact-to-rating ratio looks off. If a peer cannot talk you out of a rating, it is probably honest.
  4. Artifact re-linking. For every cell rated 3 or above, link the specific file, repo, or post that proves it. Any cell that cannot produce a link drops one notch and is flagged for the next semester of work.
  5. Year-ago comparison. Redo the grid as you would have filled it one year ago, from memory. Diff the two. The cells that moved are the direction your degree has been pulling you; that direction is real signal for concept 12.

Commit the final grid to library/raw/strength-grid.md and schedule a 12-month-later re-run on your calendar; the grid is designed to drift over time as evidence accumulates.

Transfer / How This Synthesizes Prior Semesters​

The strength grid is the point where the ten-semester curriculum gets rolled up into a usable self-assessment. Every row points at specific prior semester material:

Cells that rate high without anchoring back to a specific prior semester's artifact are usually one notch too high; the remedy is to re-rate, not to rewrite the anchor.

The grid is the most directly synthetic artifact in the module -- it is the place where the ten semesters stop being a sequence and become a map. Every subsequent cluster-4 concept and most of cluster-5 consume this map as input, so underbuilding it here will show up as thin arguments later.

See also (integrative)​

Source Backbone​

Portfolio assessment packages evidence from the whole curriculum. These books provide the technical and professional backbone for the narrative.