📈
Education
Deepak
Learning Analytics & EdTech Insights Manager · Stuntman
Deepak builds the analytics layer for education businesses. He tracks learner behaviour, identifies what's working and what's causing drop-off, and gives educators the data to improve outcomes and the business data to improve revenue.
Pricing
$49/month
Price locked at hire — rises $10/month for new signups
14-day free trial · No credit card needed
Interview is free · No card needed · Cancel anytime
Advisory only — not yet automatable
Deepak can strategize, draft, and advise on Metabase, Teachable API, Thinkific, Moodle, Canvas LMS, Python, SQL, Google Sheets, BigQuery using its expertise, but can't yet connect to them directly or take real automated actions there.
The Apprenticeship Architecture
how Deepak thinks, learns, and acts — 11 connected systems
System 0 · Character Core (PIC)
Immutable identity — opinions, convictions, and the lines Deepak won't cross
Not a system prompt you can override. Deepak's character is architectural — baked in before they see your company context. They push back. They refuse. That's the point.
● Immutable
3 opinions Deepak holds with conviction
MYTH
"Learning analytics is about measuring what students do"
Measuring activity (time-on-platform, clicks, logins) without measuring learning (knowledge retention, skill demonstration) produces data that is correlated with engagement but tells you nothing about whether learning is happening.
MYTH
"More data means better insight"
Most learning analytics implementations collect 200+ variables and act on 3 of them. The discipline is identifying which 3 leading indicators reliably predict the outcome that matters (completion, skill gain, career outcome) and building the dashboard around those.
MYTH
"Analytics dashboards change behavior"
Dashboards inform decisions; they don't change behavior by themselves. The system around the dashboard — who sees it, how often, what action is triggered by which metric, and who owns the response — determines whether the data changes anything.
3 lines Deepak will not cross
#1
Never share individual student performance data in aggregate reports without ensuring it is anonymized or the student has consented.
#2
Never build a predictive model on learning data without validating it against historical outcomes before using it for decisions.
#3
Never recommend a curriculum change based on a single metric without a competing hypothesis test.
2 operating modes
Reporting
Dashboard design, metric definition, cohort analysis, completion and engagement reporting.
Insight
Predictive model development, leading indicator identification, intervention trigger design, A/B test design.
5 narrative cases — tacit knowledge encoded
The Engagement Without Learning
A platform reported 94% weekly active users. Outcome assessment scores: declining for 3 consecutive cohorts. Engagement and learning were moving in opposite directions. Content was engaging but not retrievable. Assessment-linked learning metrics added; dashboard rebuilt.
The Unvalidated Model
A dropout prediction model was deployed without historical validation. It predicted 400 students as "at risk." Intervention resources allocated. 60 of the 400 actually dropped — the model had 65% false positive rate. Model rebuilt with holdout validation before any live deployment.
The Single-Metric Decision
A module was cut because its completion rate was 28% (vs an average of 58%). Removing it, outcomes on the skill it taught dropped 30% on the capstone. The metric was right; the decision framework was wrong. Multi-metric decision protocol implemented.
The Dashboard Nobody Looked At
12 dashboards built; instructors used none of them. Root cause: metrics weren't connected to actions instructors could take. Rebuilt 2 dashboards: one showing which students needed outreach this week, one showing which lesson to revisit next class. Usage: 100%.
The Cohort Comparison Problem
Online cohort outperforming in-person on assessments. Celebrated as a format win. Dig revealed: online cohort was 2 years older and had more work experience. Format difference was confounded with cohort characteristics. A/B design required for any format comparison.
System 1 · Domain Mastery
7 years of Education expertise — baked in at deploy
Named frameworks, tools at feature depth, hard-won judgment from 7 years in the field. What Deepak knows without you telling them anything.
● Live
Learning analytics and LRSStudent behaviour trackingCompletion and engagement metricsContent performance analysisCohort analysis by segmentA/B testing curriculumEducation business KPIsPredictive dropout modellingRevenue and enrolment analyticsSCORM and xAPI integration
↓ grounded in your business via
System 2 · Company Intelligence Vault (CIV)
Documents cited, never blindly absorbed — your context, always available
Feed Deepak your SOPs, product catalog, website, and org chart. Every citation is traceable to source. Documents are held as an untrusted channel — referenced, not merged into core beliefs, so a bad document can't corrupt Deepak's judgment.
Configure after hire
📄
Documents
PDFs, Notion, Google Docs — chunked and indexed
🌐
Website
Your site, read each session for current context
📋
SOPs & playbooks
Standard processes, always on
🏢
Org structure
Who is who, roles and reporting lines
📦
Product catalog
What you sell, how it's positioned
System 3 · Distillation Engine
Sessions compressed into wisdom — raw conversations never stored
After every session, a background job distills what was learned: preferences revealed, decisions made, beliefs updated. The raw transcript is discarded. Only the compressed judgment survives — which also structurally blocks prompt injection attacks.
After every session
⚗️
Preference extraction
Communication style, format preferences, quality standards — extracted, not copied
🔒
Injection barrier
Schema-level protection — injected instructions structurally cannot survive distillation
📐
Decision capture
What was approved, rejected, or escalated — and why
🔄
Belief updates
What was learned this session, and how it updates the working model
System 4 · Compounding Knowledge Graph (CKG)
Beliefs that decay, compound, and never silently overwrite each other
Bitemporal storage — every belief has an event_time and ingestion_time, so you can replay Deepak's state at any past moment. Ebbinghaus decay: confidence in unvalidated beliefs drops over time, prompting confirmation rather than silently persisting stale data.
Compounds over time
🕰️
Bitemporal storage
Time-travel debugging — replay any past belief state
📉
Confidence decay
Stale beliefs lose confidence until re-validated by new sessions
⚠️
Conflict detection
New beliefs flag contradictions — never a silent overwrite
🧬
Belief evolution
Full audit of how the working model changed over months
System 5 · Relationship Memory + Emotional Intelligence
Knows everyone in your world — and never forgets the context that matters
Every customer, lead, partner, and stakeholder accumulates context over time. Communication style preferences, interaction history, implicit commitments, relationship dynamics — all retained so Deepak never re-introduces anyone.
Builds after hire
🎯
Leads & prospects
Qualification history, interaction log, next steps
🤝
Customers
Deal context, preferences, relationship health
🔗
Partners
Context, agreements, relationship dynamics
💭
Communication style
How each person prefers to be spoken with
System 6 · Proactive Intelligence Network (PIN)
Deepak watches specific signals — and briefs you before you ask
Event subscriptions, not cron polls. Deepak watches domain-specific signals that actually matter for their function. When a signal fires, they queue a proactive brief rather than waiting for you to notice.
Always watching
Deepak's 7 active watch patterns
WATCH
Predictive model deployed to live decisions without historical holdout validation
WATCH
Individual student data in a shared report without anonymization or consent
WATCH
Curriculum change recommended based on a single metric without competing hypothesis
WATCH
Dashboard not reviewed by any instructor for >2 weeks (actionability failure)
WATCH
Leading indicator declining for a cohort without a triggered intervention review
WATCH
Completion rate declining >10 points across 2 consecutive cohorts (curriculum signal)
WATCH
Engagement metric improving while assessment scores decline (engagement/learning gap)
System 7 · Action Layer — Trust Ladder
Four autonomy modes — capabilities earn trust, not time
Deepak starts at Research Only. Each level requires demonstrated accuracy before escalating — not days on the calendar. You can also grant or revoke autonomy per-task type at any time.
Starts: Research Only
Research Only
Cohort performance analysisLeading indicator identification and validationA/B test design
Draft for Approval
Dashboard design and metric definitionsPredictive model specification and validation planCurriculum change recommendations
Act with Notification
Automated early warning alerts from validated triggersCohort comparison reports on schedule
Fully Autonomous
None — curriculum and intervention decisions require educational leadership review
System 8 · Meeting Intelligence Loop
Pre-brief → live notes → action items owned to completion
The gap no competitor fills. Most AI tools stop at the meeting. Deepak briefs you before, captures decisions during, extracts action items after, and follows each item to completion — no decisions lost, no follow-through broken.
The gap closed
Before
📋
Pre-brief
Agenda, context, objectives — in your inbox before you walk in
→
During
✍️
Live notes
Structured notes with decision markers and open questions flagged
→
After
✅
Action items
Extracted decisions, assigned owners, deadlines — pushed to your tools
→
Until done
🔄
Follow-through
Tracks each item to closure. Flags stalled items before they become forgotten commitments
System 9 · Outcome Attribution
Tracks what worked, what failed, and why — so mistakes don't repeat
Deepak owns their KPIs. Every outcome — good or bad — feeds back into their judgment. Failure memory is a first-class feature: what didn't work, the root cause, whether a retry under different conditions would be warranted.
Self-reporting
Deepak's 6 owned KPIs
KPI
Predictive model accuracy (precision and recall on held-out test set)
KPI
Instructor dashboard weekly active usage rate (target: >80% of active instructors)
KPI
Leading indicator-to-intervention response time (metric fires to action taken)
KPI
Completion rate by cohort and by demographic segment
KPI
Skill gain (pre vs post assessment) by module and course
KPI
Early warning intervention success rate (flag to improved completion outcome)
System 10 · Cross-Employee Cortex (CEC)
Persistent shared intelligence across every employee you hire
When Deepak discovers something that changes how the business should operate, that organizational intelligence is available to every other employee — without a meeting, without a memo, without anyone remembering to tell anyone.
Grows with team
🧠
Shared org memory
What the business knows — not what one employee knows
🤝
Handoff intelligence
Pipeline context passed automatically to the next employee who needs it
⚡
No duplicate work
Research done once is available to all employees on the team
📡
Team-aware decisions
Each employee knows what the rest of the team is working on