Ravi Menon
BRD v0.9 is ready for review. Comments are due Mon 8 Mar; approval is targeted for Fri 12 Mar.
YouWrite the review request, then work through 61 comments.Week 3
Operator ProgramCohort 1 applications openApply
Become a job-ready project manager in 26 weeks for $0
Free from start to finish, with nothing to pay back. The Operator Program is Gantt Guild's live, 26-week training in project and product management. Run four projects through a full fiscal year, write every email, purchase order and Slack message yourself, and meet employers at three Career Days.
47 tools, 33 from GitLab's handbook.
Your own accounts.
3
Career Days, weeks 9, 17, 26.
Meet employers online.
129
Real deliverables,
reviewed live every Friday.
Your inbox, FY28
1 / 10
Ten moments from the year you'll run. The company, people and numbers are the fictional cases in Projects, For Real, set against GitLab's public handbook.
[ 47 tools · your own accounts ]
GitLabGitLab Ultimate trialGitLab API + glab CLIMermaid diagramsGitLab DuoSlackZoomGeekbotLoomGoogle CalendarCalendlyGoogle DocsGoogle SheetsGoogle SlidesGoogle FormsFigmaFigJamMuralDovetailZipCoupaDocuSignKantataWorkday
OktaSalesforceClari and GongGainsightZendeskCulture Ampincident.ioPagerDutyGrafanaDuckDB / SQLitePython + JupyterTableau PublicJiraJira Product DiscoveryJira Service ManagementConfluenceMicrosoft ProjectMicrosoft Planner + TeamsExcel + Power BISmartsheetNotionClaude Code (Claude Pro)Codex (ChatGPT Plus)
[ What you get ]
Not a course. A year on the job in six months.
Everything a first-year project manager learns the hard way, compressed into 26 live weeks, with a review every Friday.
Tuition
$0
The whole program, not just the application. No deposit, no income-share agreement, and nothing owed if you leave or get hired.
- Six months of live training
- Friday reviews, one to one
- The 1,750-page book
- Three Career Days
The case
4 projects through one fiscal year
Ledger, Beacon, Run the Engine and Harbor, all competing for the same people and money.
Your portfolio
129
deliverables you write yourself: charters, RFPs, purchase requests, schedules, change orders, specs and queries.
Live, four days a week
- MonKickoff90 min
- WedCode lab90 min
- ThuBusiness case60 min
- FriReview2 hrs
About 6 live hours, all recorded. 20 to 25 hours a week in total.
Code from week one
SELECT type, AVG(inside_85) FROM backtests GROUP BY type;
Python from week 1, SQL from week 9. No experience needed.
Career Days
3 chances to get hired
Week 9Week 17Week 26
Hired in week 9? Take the job. You'll never owe us anything.
Your AI agent, supervised
An agent drafts. You check every number.
Claude Code or Codex, about $20 a month, paid to Anthropic or OpenAI. It's the one cost in the program, and we never see it.
YouCompute Ledger's estimate at completion from cost_ledger.csv.
AgentCPI 0.97, so EAC = BAC ÷ CPI = $619k.
YouThe overrun was partly a one-off. Show all four methods and the bottom-up estimate.
Reported: EAC $625k, range $619k to $632k, method stated.
[ The plan ]
From day one to Operator in five stages
Every stage ends with a live checkpoint. Meet the standard and you move up a rank. Miss it, and you get coached and try again within a week.
Plan four projects to a steering-committee standard
Learn how a 2,000-person company is wired, then plan all four projects: the business case, the charter, the RFP, the purchase request, and a schedule you can defend.
You'll be able to
- Write a business case, a charter and an RFP
- Build a critical-path schedule and simulate it 20,000 times
- Walk a $1.14M purchase through five approvals
- Answer your own questions in Python
Week 8
Phase review: defend your integrated plan live. Pass, or revise and present again within a week.
Run the year when things go wrong
The projects go live. A vendor's connector fails, a production incident hits 12% of page views, and a bank buys a subsidiary. You run sprints and flow, report honestly and control change.
You'll be able to
- Run Scrum and Kanban with real flow metrics
- Report earned value a sponsor believes
- Lead a live incident and write the blameless review
- Price a change order on a fixed-price contract
Week 15
Phase review: a closure report, and a retrospective you facilitate live.
Mini Career Day in week 9
Decide what's worth building
Follow Beacon from twelve customer interviews to general availability: vision, OKRs, discovery, prioritization, specs, design, metrics, experiments, pricing and launch.
You'll be able to
- Run customer interviews and find the pattern
- Set OKRs and score a backlog with RICE
- Write a spec and a metrics tree
- Design an experiment with a decision rule
Week 21
Phase review: a product decision presented and defended, plus the data exercise.
Mini Career Day in week 17
Rebuild it in the tools employers run
Your capstone, rebuilt in GitLab in depth, Jira and Confluence, Microsoft Project, and a spreadsheet someone else can audit. Then you automate the boring parts.
You'll be able to
- Run the same project in GitLab and in Jira
- Schedule in Microsoft Project with a baseline
- Build a workbook anyone can audit in 30 seconds
- Ship a working automation with your agent
Week 24
Automations demoed live, including what happens when they fail.
Show employers what you can do
No new reading. You publish your portfolio, sit a timed 160-question practice exam, rehearse interviews with working project managers, and present at Career Day.
You'll be able to
- Publish a portfolio employers can open
- Sit a timed 160-question practice exam
- Interview with real work to point to
- Present your work at Career Day
Week 26
Career Day: online, with every employer in our network.
Keep growing after you're hired
If your employer chooses Graduate Support, your training continues for another 6 to 12 months while you work. Either way, your portfolio and the book are yours to keep.
You'll be able to
- Weekly training alongside your new job
- Monthly coaching on your employer's tools
- Manager check-ins with a written progress note
- Your portfolio and the book, yours for good
Optional
Graduate Support is chosen and paid for by your employer. It never costs you anything.
1 / 6
[ Skills ]
What you'll be able to do, stage by stage
Pick a point in the program to see where each skill stands. Levels are what our reviewers expect to see in your work by then.
- Planning and schedulingCharters, WBS, critical path, Monte CarloCan lead it
- MoneyBudgets, earned value, procurement, change ordersCan lead it
- Agile and flowScrum, Kanban, WIP limits, flow metricsCan lead it
- Risk, change and incidentsRegisters, change control, incident commandCan lead it
- Stakeholders and writingRequests, status, escalations, steering packsCan lead it
- ProductInterviews, OKRs, RICE, specs, experimentsCan lead it
- Data: SQL and Pythonpandas, joins, funnels, cohorts, simulationCan lead it
- Supervising AIDelegating to an agent and checking its workCan lead it
- The businessStrategy, accounting, finance, marketing, operationsCan lead it
- ToolsGitLab, Jira, Microsoft Project, Sheets, SlackCan lead it
- 1Aware
- 2Guided
- 3Independent
- 4Confident
- 5Can lead it
[ The work ]
Every email. Every purchase order. Every Slack message. You write all of it
You don't answer quiz questions about a charter. You write the charter, and every other document a sponsor would sign, then defend it on Friday.
01Week 3Ledger
Ask for a review people can act on
The requirements document needs sign-off from Tax, Payroll and every other affected team. One message tells each reviewer their section, the deadline and the approval date.
The standardCould a newcomer act on it without asking a single question?
Slack message
#proj-ledger
You
YouMon 1 Mar, 10:02
BRD v0.9 is ready for review. Please focus on your area: Tax §5.3, Payroll §5.4, EBAs §5.6.
Comments by Mon 8 Mar. Approval targeted for Fri 12 Mar.
61 comments, 58 resolved
02Week 8Ledger
Buy a $1.14M platform, properly
A three-year contract has to clear sanctions screening, security, privacy, legal and negotiation before FP&A, the VP of IT and the CFO approve it. You write the request and walk it through every gate.
The standardEvery approver can see the need, the evidence and the money on one page.
Purchase request
Ledger expense platformReleased 4 Jun
Spendwise, Inc.
- Term
- 3 years at $380,000 a year
- Total
- $1,140,000
Reviews
- Sanctions
- Security
- Privacy
- Legal
- Negotiation
Approvals
- FP&A
- VP IT
- CFO
- DocuSign
03Week 7Ledger
Check the date before you promise it
The plan says 82 days. You model every activity as a range, run the schedule 20,000 times in Python, and find the plan's own date comes true 37% of the time. Then you recommend a date you can defend.
The standardNo date without a network and a probability behind it.
Python
$ python ledger_monte_carlo.py
- P(≤ 82 days)
- 37%
- P50
- 83.2
- P80
- 86.6
- P95
- 90.2
04Week 14Run the Engine
Lead an incident, live
At 13:40 a feature flag reaches 100% of groups and epic boards start timing out. You run the incident channel, mitigate, then write the blameless review and its corrective actions.
The standardYou're the incident lead for 30 minutes, in a live simulation.
Incident
INC-2219Severity 2Resolved
Epic boards timing out for large groups
- Feature flag enabled for 100% of groups
- Incident declared
- Flag disabled, service mitigated
- Resolved; blameless review required
05Week 14Harbor
Say yes to more scope, at a price
The bank buys a subsidiary and wants 2,000 more repositories moved before its change freeze. On a fixed-price contract, you estimate the work, price it and write change order CO-1.
The standardScope, estimate, price and approvals in writing before any work starts.
Change order
Change order CO-1Signed
Meridian Bank, SOW-4471: subsidiary repositories
- +2,000repositories, as wave 3b
- +6 weeksnew end date 27 December
- $96,000fixed fee, on milestones M3 and M4
- 450 hourswith the basis of estimate
06Week 19Beacon
Find what the average hid
Forecast accuracy is flat at 82%. You query it by forecast type and find epics improved nine points while the mix shifted toward them. Report by type, never only blended.
The standardEvery number comes with the query that produced it.
SQL
SELECT quarter, type, ROUND(100.0 * AVG(inside_85)) AS pct FROM forecast_backtests GROUP BY quarter, type;
| type | Q3 | Q4 |
|---|---|---|
| milestone | 84 | 85 |
| epic | 71 | 80 |
| all | 82 | 82 |
[ The case ]
One company. Four projects. One fiscal year
Real companies run many projects at once, competing for the same people, money and executive attention. You run all four, week by week, from February 2027 to January 2028, inside an all-remote software company modeled on GitLab's public handbook.
LedgerInternal program, plan-driven
Replace the company's expense system
An IT program to move 2,000 employees off an aging expense tool that takes eleven days to reimburse, before its support ends on 31 December. Sarbanes-Oxley controls apply to everything.
- Money
- $600k to build ($695k with contingency) and a $1.14M three-year platform contract
- Dates
- RFP in March, contract on 4 June, go-live Monday 4 October, hypercare to 5 November
- Cast
- Priya Raman (program manager), Daniel Okafor (sponsor), Sofia Alvarez (business lead), Chen Wu (integrations)
Ledger in FY28
RFP and contractBuild and testGo-liveClose
What goes wrong
The vendor's single sign-on connector fails testing on 11 August. Third-quarter spending runs $40k over Finance's allocation. Card feeds from 22 providers threaten the date.
Some of what you write
- Business case
- Charter
- BRD review
- RFP scoring
- Purchase request
- Schedule network
- Monte Carlo
- Earned value
- Steering packs
- Cutover plan
- Final report
BeaconProduct development, monthly releases
Ship a forecasting feature inside GitLab
A product team builds delivery forecasts (“85% likely by 5 November”) into GitLab's milestones and epics, from twelve customer interviews to general availability.
- Money
- No budget line: about nine person-months of team capacity
- Dates
- Discovery February to April, releases 20.1 to 20.3, general availability in 20.4 on 16 September
- Cast
- Elena Petrova (senior product manager), Sam Adeyemi (engineering manager), Alex Kovalenko and Fatima Zahra (engineers)
Beacon in FY28
DiscoveryBuild, releases 20.1–20.3GAImprove
What goes wrong
A security fix and a production incident take about 30% of the team's capacity in August. The team's own forecast says GA will miss, so scope is cut and the date holds.
Some of what you write
- Opportunity canvas
- Interview guide
- OKRs
- RICE scoring
- Spec
- Metrics tree
- Experiment brief
- Release post
- Decision memo
Run the EngineOngoing operations, Kanban and flow
Keep a product area healthy
The same engineers who build Beacon also own bugs, security fixes and performance for a large product area. Work that never ends needs flow, not sprints.
- Money
- No project budget; service levels and an error budget instead
- Dates
- Every day, with a monthly release on the third Thursday
- Cast
- The same group as Beacon, with on-call engineers and an incident lead
Run the Engine in FY28
Triage, fixes and monthly releases, all year
What goes wrong
A feature flag reaches 100% of groups at 13:40 UTC on 3 August. Within 48 minutes, 12% of epic board views are failing. The month's error budget ends 96% spent.
Some of what you write
- Triage policy
- Kanban with WIP limits
- Flow metrics
- Rollout issue
- Incident channel
- Blameless review
- Corrective actions
- Monthly indicators
HarborCustomer engagement, fixed price, agile delivery
Deliver a fixed-price migration for a bank
Professional services moves Meridian Bank's 3,000 developers to a dedicated GitLab instance: 8,000 repositories, 40 pipelines and training, for a fixed fee of $420,000.
- Money
- 1,980 planned hours against a 45% margin target
- Dates
- Kickoff 20 July, change order 21 September, the bank's change freeze from 20 November
- Cast
- Diego Martins (project manager), Rachel Osei (technical architect), Priyanka Shah (the bank's head of platform engineering)
Harbor in FY28
Iterations and migration wavesBank change freeze
What goes wrong
The bank buys a subsidiary and wants 2,000 more repositories moved before the freeze. In a customer review, its platform lead projects a February finish.
Some of what you write
- SOW review
- Engagement charter
- Iteration reviews
- Change order
- Capacity forecast
- Customer status
- Acceptance records
- CSAT survey
[ How we train ]
Not lectured like a webinar. Trained like special forces
High standards, real conditions and live review every week. Nobody gets carried, and everyone who wants to finish gets the coaching to do it.
- 01
Real artifacts, every week
You don't answer quiz questions about a charter. You write the charter, the RFP, the purchase request and the status report, to the standard a sponsor would sign.
129 deliverables - 02
Reviewed live, every Friday
Your work goes on the screen in front of the cohort and gets challenged against that week's written standard. Meet it, or revise and show it again.
26 reviews, 3 phase reviews - 03
Things go wrong on schedule
A vendor's connector fails. A feature flag breaks a page for 12% of users. A bank buys a subsidiary mid-project. You handle each one as it happens.
The book's failures, run live - 04
Your accounts, your tools
You create your own accounts in the tools GitLab's handbook documents, on free plans and trials. Nothing is set up for you, just like a real first week.
47 tools, 1 paid - 05
Code from week one
Python from day one and SQL from week 9, on a real database engine on your own laptop. By the end you query, simulate and automate your own reports.
A lab every week - 06
Check the machine
Every week you put an AI agent to work, then verify what it produced: the query, the schedule, the contract clause. Your judgment is what you're trained for.
26 agent labs
[ Hands on ]
Yes, you'll write the SQL
Project managers who can answer their own questions get trusted with bigger ones. Every week has a lab, built on the four projects' own data.
- Python from week 1. Notebooks, pandas, simulation and automation, starting from zero.
- SQL from week 9. DuckDB on your own laptop: joins, funnels, cohorts and the traps in each.
- An AI agent every week. Claude Code or Codex drafts; you verify every number before it goes anywhere.
No coding background needed. The labs start from installing Python.
accuracy_by_type.sql
SELECT quarter, type,
SUM(inside_85) AS inside,
COUNT(*) AS total,
ROUND(100.0 * AVG(inside_85), 0) AS accuracy_pct
FROM forecast_backtests
GROUP BY 1, 2
ORDER BY 1, 2 DESC;result
quarter | type | inside | total | accuracy_pct
Q3 | milestone | 420 | 500 | 84
Q3 | epic | 71 | 100 | 71
Q4 | milestone | 170 | 200 | 85
Q4 | epic | 320 | 400 | 80ReviewOverall accuracy is 82% in both quarters, yet epics improved nine points. The mix shifted toward epics. Report by type, never only blended.
ledger_monte_carlo.py
def beta_pert(o, m, p):
a = 1 + 4 * (m - o) / (p - o)
b = 1 + 4 * (p - m) / (p - o)
return o + rng.beta(a, b, RUNS) * (p - o)
finish = {}
for act, (estimate, preds) in ACTIVITIES.items():
start = np.max([finish[p] for p in preds], axis=0) if preds else 0
finish[act] = start + beta_pert(*estimate)
print(f"P(<= 82 days) {np.mean(finish['Q'] <= 82):.0%}")output
P(<= 82 days) 37%
P50 83.2 days P80 86.6 days P95 90.2 daysReviewThe plan's own date comes true in 37% of 20,000 runs. Commit to a date with a fallback, and watch the card feeds.
YouCompute Ledger's earned value by control account from cost_ledger.csv and give me the estimate at completion.
AgentCPI 0.97, SPI 0.95. EAC = BAC ÷ CPI = $619k.
YouThe integrations overrun was partly a one-off vendor defect. Show all four EAC methods, plus the bottom-up estimate from the control account owners.
AgentBAC ÷ CPI $619k. AC + (BAC − EV) $611k. With CPI × SPI $632k. Bottom-up $625k.
Your reportEAC $625k (bottom-up), range $619k to $632k, with the method and the reason stated. The agent's first answer was arithmetic; the judgment was yours.
[ The business ]
And the business around every project
The project managers who get the hard projects understand the business. Alongside the four projects you work through the core of a first-year MBA: 58 chapters in seven subjects, each with a case and an exercise. They're the subjects at the heart of required curricula like Harvard Business School's.
- Strategy8 chapters · weeks 9–10
Run the Five Forces on Larkspur, a $140M freight-software company, then decide whether to build, buy or partner.
- Financial accounting8 chapters · weeks 11–13
Take a software bundle through the five revenue-recognition steps; decide what Ledger capitalizes.
- Finance9 chapters · weeks 13–15
Compute Larkspur's cost of capital, then value the option to expand Beacon.
- Marketing11 chapters · weeks 16–19
Segment an outdoor brand's customers, price a new tent line and design a channel strategy.
- Operations9 chapters · weeks 19–21
Apply Little's law to Harbor's migration pipeline; forecast demand and set order quantities.
- Leadership and organizational behavior6 chapters · weeks 22–23
Redesign a failing global team and prepare a negotiation.
- Entrepreneurship7 chapters · weeks 23–24
Write a start-up's riskiest hypotheses, design the cheapest tests, and pitch its seed round.
Gantt Guild is not affiliated with, or endorsed by, Harvard University or Harvard Business School.
[ Your own accounts ]
GitLab's own tools. Your own accounts
GitLab documents how it works in a public handbook, right down to the tools. You use the same ones, on free plans, developer accounts and trials you create yourself. Enterprise-only systems like Coupa and Workday are studied from real documents, then reproduced in free tools.
Work, code and the handbook
GitLabGitLab Ultimate trialGitLab API + glab CLIMermaid diagramsGitLab Duo
Communication
SlackZoomGeekbotLoomGoogle CalendarCalendly
Documents and models
Google DocsGoogle SheetsGoogle SlidesGoogle Forms
Design and research
FigmaFigJamMuralDovetail
Money, contracts and people
ZipCoupaDocuSignKantataWorkdayOkta
Customers and revenue
SalesforceClari and GongGainsightZendeskCulture Amp
Incidents and reliability
incident.ioPagerDutyGrafana
Data and analytics
DuckDB / SQLitePython + JupyterTableau Public
The other shops
JiraJira Product DiscoveryJira Service ManagementConfluenceMicrosoft ProjectMicrosoft Planner + TeamsExcel + Power BISmartsheetNotion
Your AI agent (pick one)
Claude Code (Claude Pro)Codex (ChatGPT Plus)
Your total
~$20/month
For one AI agent, Claude Pro or ChatGPT Plus, paid to Anthropic or OpenAI. Everything else is a free plan, a developer account or a trial.
- Free plan
- Free dev account
- Free trial
- Walkthrough
- ~$20/mo
[ What it costs ]
Free means the whole program
Not a free trial, a free first module or a free application. Six months of live training, feedback and Career Days cost you nothing, now or later.
- No tuition, no deposit, no deferred tuition
- No income-share agreement, ever
- Leave any time, for any reason, and owe nothing
Employers fund it. A company that hires a graduate it met through us pays us a flat placement fee once they start, and some pay us to keep developing the person they hired. The fee never comes out of your pay, and it only works if we train people well.
StatementOperator Program, Cohort 1
Bill to: you
| 26 weeks of live training | $0.00 |
| Friday reviews and one-to-one feedback | $0.00 |
| Projects, For Real, all 1,750 pages | $0.00 |
| Three Career Days with employers | $0.00 |
| Income-share agreement | None |
| If you leave early | $0.00 |
| If you get hired | $0.00 |
| Total due to Gantt Guild | $0.00 |
Your one outside cost: an AI coding agent for the weekly labs, about $20 a month (Claude Pro or ChatGPT Plus), paid to the provider, never to us. About $120 over six months.
[ Who it's for ]
Is this for you?
People with a college degree or equivalent experience who want to run projects for a living. Career returners, career changers and people who left college are all welcome.
A good fit if
- You have a college degree, or equivalent experience
- You're starting out, coming back from a break, or changing careers
- You can give 20 to 25 hours a week for 26 weeks
- You live in the US and are authorized to work there
- You'd rather be reviewed honestly than praised vaguely
Not a fit if
- You want a certificate without doing the work
- You can't give 20 hours a week right now
- You need visa sponsorship
- You're looking for self-paced videos
- You want a job guarantee. Nobody honest offers one
Recent graduates
A degree, and job posts that want three years of experience. Leave with a year of real work to show.
Career returners
Back after caring for family, illness, service or a move. Get current on the tools and on AI, fast.
Career changers
Teachers, nurses, store managers, engineers: you already run things. Learn the language and the tools.
Self-taught and early leavers
No degree? Equivalent experience counts, and so does everything you've organized and delivered.
[ Career Days ]
Hired before you finish? Good
Employers meet the cohort online in weeks 9, 17 and 26. If you get an offer in week 9, take it. Leave with nothing to pay back, or keep attending if your new job allows.
15913172126
- Week 9Mini Career Day
Integrated project plans, defended live.
- Week 17Mini Career Day
Sprints, earned value, an incident review and a closed project.
- Week 26Career Day
Complete portfolios, tool builds and an AI playbook.
For employers
Meet the cohort free. Pay only when a hire starts.
Career Days cost nothing to attend. When someone you met through us joins your team, you pay a flat $5,000, not a percentage of salary. Graduate Support adds 6 to 12 months of training, from $8,000. Graduates never pay, and nothing limits where they can work.
[ Compare ]
How it compares
| Online certificate courses | Coding bootcamps | Full MBA programs | Gantt Guild | |
|---|---|---|---|---|
| Cost to you | Subscription or course fees | Tuition, or a share of future pay | Tuition, often six figures | $0. About $20 a month for your own AI agent |
| Format | Self-paced video and quizzes | Live, often full-time | Live cases, one to two years | Live online, 20 to 25 hours a week, 26 weeks |
| What you produce | Quiz scores and a certificate | Software projects | Case discussions and papers | 129 project artifacts across four projects |
| Real workplace tools | Rarely | Developer tools | Rarely | 47 tools, many from GitLab's handbook |
| Code | None or optional | The focus | Rarely | Python, SQL and an AI agent, weekly |
| Business core | No | No | Yes | Seven MBA core subjects |
| Live review of your work | Automated or peer-graded | Yes | Yes | Every Friday |
| Path to a job | Job boards | Career services | Recruiting season | Three Career Days with hiring employers |
Comparisons describe typical programs in each category, not any specific provider.
[ Your instructor ]
Taught by the person who wrote the book
Mohit Bhardwaj is the founder and Head of Engineering at Dart Browser and the author of Projects, For Real. He teaches every live class himself.
The book follows four projects through a full fiscal year inside one company and shows where they collide: over the same people, the same budget cycle and the same executives' attention. It explains the theory, then shows exactly what the documents, meetings and decisions look like. It's released openly, and it's the program's textbook.
- 152chapters
- 1,750pages
- 306exhibits
- 34templates
- 160practice questions
[ From the blog ]
How projects actually run
- Incidents
INC-2219: anatomy of a production incident, minute by minute
A feature flag reached 100% of a large SaaS platform at 13:40 UTC. Forty-eight minutes later, 12% of epic board views were failing. What happened, why four layers of defense lined up, and what a project manager should do in the room and after it.
- Procurement
Five weeks from “yes” to signature: how a public company actually buys software
The evaluation team picked a vendor on 30 April. The contract was signed on 4 June. In between: sanctions screening, security, privacy, legal, negotiation, FP&A and a three-level approval chain. The path, the documents and the mistakes that cost a week.
- Control
“$7k under budget” was wrong: earned value on a real system rollout, month by month
At the end of August, Ledger had spent $363k against $370k planned. A budget report would call that good news. Earned value showed the project behind schedule and over cost, pinned the cause to one control account, and forecast the final cost within 1%.
[ FAQ ]
Questions, answered
Can't find yours? A person reads every email.
Is the whole program really free?
Yes. No tuition, no deposit, no income-share agreement, nothing owed if you leave or get hired. The only outside cost is an AI agent subscription of about $20 a month, which you pay to Anthropic or OpenAI directly.
Who pays for it, then?
Employers. A company that hires a graduate it met through us pays us a flat fee once they start. It never comes out of your pay, it never limits where you can apply or work, and you never need our permission to take, turn down or leave any job.
Do you give us accounts for the tools?
No. You create your own accounts, mostly on free plans, plus free developer accounts and trials in the weeks that need them. Setting up your own workspace is part of the training.
How much time does it take?
20 to 25 hours a week for 26 weeks, with about 6 hours of live sessions you can catch up on from recordings.
Do I need to know how to code?
No. Python starts from installation in week 1 and SQL in week 9.
Who can apply?
Anyone 18 or older who lives in the US and is authorized to work there, with a college degree or equivalent experience. Career returners, career changers and people who left college are all welcome.
What does “Operator” mean?
It's what we call graduates of the Operator Program: people who can run real work end to end, from the business case to the closing report. You earn a rank at each stage: Planner, Project Lead, Product Lead, Specialist, then Operator.
[ Cohort 1 · applications open ]
It's free.
Become an Operator
Applying takes about ten minutes. A person reads every application, and you'll hear from us within 7 days.
FY28
Ledger
IntakeRFP and contractBuild and testGo-liveHypercareClose
Beacon
DiscoveryBuild, releases 20.1–20.3GAImprove
Run the Engine
Triage, fixes and monthly releases, all year
Harbor
SOWIterations and migration wavesBank change freeze