Job search
Codesmith Immersive: Check the Current Format, Project Work, and Enrollment Terms
Check current Codesmith immersive format, project expectations, tuition and enrollment uncertainties separately from other programs.
Codesmith currently presents its Software Engineering + AI/ML Immersive as a live remote program costing $22,500, with 14-week full-time and 39-week part-time formats. Timing remains unresolved: the current page anticipates a fall cohort only once client timelines are confirmed. Applications being open does not establish a confirmed start date or a reserved place.
This guide checks public pages and the currently linked syllabus on October 23, 2026. It separates the paid immersive from free NYC programs and shorter AI training, then identifies what remains unresolved before enrollment. Start with recent engineering roles on LandOffer and the responsibilities you need to demonstrate; identify the gap in your own work before comparing program formats and costs.

The current offer is clearer than the next start date
The old immersive URL now redirects to Codesmith's For Individuals page. That page is the current starting point for format, tuition and availability. It describes advanced JavaScript and TypeScript, full-stack projects, AI/ML work, open-source products and hiring support.
Its timing notice connects future immersive cohorts to enterprise and government client programs. It anticipates a fall cohort when those timelines are confirmed, without publishing a firm immersive date in that section. Do not turn that statement into a particular November start, an internship commitment or a client-funded job for every graduate.
Search results still expose older versions with 2025 cohort dates, detailed installment schedules and admissions steps. Those are useful research leads, not an offer available today. A reader planning leave from work needs a confirmed cohort, cancellation terms and contingency arrangements before rearranging employment or childcare.
The current page publishes 14 weeks full-time or 39 weeks part-time, with the latter held in evenings and on Saturdays. Exact current weekly hours, time zone, attendance rules and independent-work expectations need confirmation. Live remote teaching still requires simultaneous availability; it is not interchangeable with completing recorded lessons whenever convenient.
Read the linked syllabus with its visible date
The currently linked 12-page immersive syllabus labels its latest curriculum update August 2024. I read all twelve pages and inspected selected pages visually. The current page links that document, but the document itself is dated 2024. That distinction matters when using its timetable or admissions description.
The syllabus describes core lectures followed by pair programming, individual work, team builds, improvements to another group's codebase and a larger open-source product. These categories help you ask what engineering work you would own. They do not establish that today's next cohort follows every week number printed in that document.
Its technical coverage includes TypeScript, React and state management, Node, relational and nonrelational databases, authentication and testing. The AI material describes integrating models, retrieval-augmented generation, fine-tuning, evaluation and deployment considerations. A prospective student should ask which parts are required now, how model and infrastructure costs are handled, and what assessment distinguishes a working demonstration from a reliable system.
The older document places a Python machine-learning unit after graduation as optional. Do not use that section to imply all graduates complete a deep machine-learning specialization. The current public overview emphasizes applied AI work within software engineering; verify the precise curriculum version and required deliverables in the next cohort's documents.
Ask what project ownership actually produces
Codesmith's current overview describes solo, scratch, iteration and AI/ML projects, followed by guided open-source developer-tool work. Pay particular attention to iteration: improving an unfamiliar repository tests a different skill from starting a fresh application with your preferred framework.
For each project, ask who scopes the problem, how teams are formed, what instructors review and what happens when a teammate leaves. Clarify which design decisions, tests, documentation and operational constraints instructors review, and whether a final demonstration is a separate requirement. Those answers tell you what feedback and practice the tuition buys; a framework list cannot establish that scope.
Make your own contribution visible in the portfolio record. Save a small design decision, the change you implemented, its tests and the limitations you found. For a team product, distinguish your work from infrastructure supplied by instructors or code written by partners. Use the actual course and repository context when describing it to employers; an educational build is not automatically paid professional experience.
AI projects need equally precise evidence. A successful prompt on one example does not establish retrieval quality, safe handling of missing evidence or reliable behavior after a dependency changes. An evaluation artifact helps a reviewer see both your reasoning and the limits of the result. The following completed example shows the kind of reasoning to seek, without pretending to reproduce a Codesmith assignment.
Worked asset: an original project-evaluation artifact
Nadia's release-note assistant is fictional. She wants to build a small tool that answers questions from an approved runbook. This is an independent portfolio planning example, not a resident project, an instructor assessment or evidence of admission.
Her completed design note separates three responsibilities. A database stores versioned source records; a retrieval adapter returns a record identifier and supported answer; the interface displays that source or a clear no-evidence response. She chooses explicit source references because a plausible answer without a supporting record would be unsuitable for a deployment decision.
For a safe local exercise, she uses invented records about deployment ownership, a build artifact and rollback prerequisites. The fixture contains no employer information. The prototype is a deterministic lookup adapter, not an embedding service or language model. She writes four evaluation cases before considering model integration.
| Local case | Expected result | Initial adapter | Revised adapter |
|---|---|---|---|
| Canonical deployment-owner query | Release Engineering, record D2 | Pass | Pass |
| Same query in uppercase | Same source-backed answer | Fail | Pass |
| Rollback query with surrounding spaces | Previous manifest, record D3 | Fail | Pass |
| Unlisted production-capacity query | No evidence; no invented number | Pass | Pass |
The initial implementation matched exact strings. The revision strips surrounding query whitespace and converts query text to lowercase before matching; unsupported queries still return no evidence. The accompanying offline fixture and script were executed: the initial exact-match adapter passed two cases and failed two; the revised adapter passed all four. Those counts describe only these four synthetic cases, not production accuracy or Codesmith's teaching results.
Nadia's completed review decision is to retain the source identifier and fallback behavior, add a version field to any future API response, and postpone claims about AI reliability. A later model-based version would need separate retrieval and answer-quality tests, access controls, realistic failure cases and evaluation on data she is authorized to use.
This artifact gives her a concrete project question for Codesmith: would the current curriculum review that separation of responsibilities and the missing tests, or only whether a demo runs? The question remains unsent. No model, cloud deployment, user upload or paid learning system was used for this example.
Keep the other Codesmith programs separate
The current individuals page also lists AI & ML Technical Leadership at $4,595, or $895 for immersive alumni. It describes four weeks of live online study, with weekday-evening and Saturday sessions, aimed at experienced engineers, data scientists and technical managers. That is a separate product with a different audience; the alumni rate is not a discount available to every new immersive applicant.
Future Code NYC is another separate route: a 42-week live online program with no tuition for eligible participants. The current overview lists 40 weekly hours and November 30, 2026–September 17, 2027. Its eligibility includes NYC residency and additional work, education, income and availability conditions. These dates and funding rules must not be copied onto the paid immersive.
Foundational Launchpad is primarily asynchronous and no-cost for eligible participants. Its current options are approximately 20 weekly hours over eleven weeks or six weekly hours over 36 weeks, with optional study and pairing sessions. The page lists November 6, 2026 as the application deadline. Its optional Future Code pathway involves separate consideration, not automatic admission.
Public preparation also has access boundaries. I inspected the CSX catalog; following its Recursion unit redirected to login, where I stopped. Codesmith's separate public recursion tutorial was readable and discusses base cases, call-stack growth and tracing execution. Reading that article is not a trial of instructor feedback, the immersive curriculum or a completed CSX challenge.

Worked asset: format compatibility and an enrollment budget
Beck's planning record is independent of Nadia's project. Beck has a $32,000 transition reserve and works weekday daytime hours. He can protect Monday–Thursday evenings and Saturday daytime, totaling 26 weekly hours including independent study. He has neither a confirmed start date nor a current enrollment offer.
The August 2024 syllabus describes a part-time timetable of three hours Monday–Thursday and six on Saturday: 18 scheduled hours weekly. Its full-time timetable spans ten hours on each weekday plus 7.5 on Saturday: 57.5 hours across those blocks. These are historical scheduled spans, not verified current teaching hours, and they may include breaks.
| Format being checked | Current published length | Provisional comparison with Beck's calendar |
|---|---|---|
| Full-time immersive | 14 weeks | Older daytime blocks conflict with work; the 57.5-hour span exceeds his 26-hour availability by 31.5 |
| Part-time immersive | 39 weeks | Older 18-hour evening/Saturday pattern leaves eight of his 26 hours for other work; current schedule still unconfirmed |
Beck's completed calendar decision is to rule out combining his existing job with the older full-time pattern. Part-time remains a possibility, not a verified fit: he needs current live hours, time zone and expected project effort before deciding that eight remaining hours are enough. His reserve calculation cannot solve a timetable conflict.
He then compares the current $22,500 tuition with an assumed $250 weekly transition allowance. The allowance is his invented planning input, not Codesmith's fee or a full living-cost estimate. It excludes lost earnings, taxes, loan interest, equipment and any separately charged services.
| Beck's scenario | Tuition plus assumed allowance | Remaining from $32,000 |
|---|---|---|
| Fourteen-week plan | $22,500 + (14 × $250) = $26,000 | $6,000 |
| Thirty-nine-week plan | $22,500 + (39 × $250) = $32,250 | −$250 |
| Fourteen weeks plus eight-week delay reserve | $26,000 + $2,000 = $28,000 | $4,000 |
| Thirty-nine weeks plus eight-week delay reserve | $32,250 + $2,000 = $34,250 | −$2,250 |
These calculations were executed locally. Eight weeks is a sensitivity scenario, not a forecast of Codesmith's delay or a promised deferral right. Beck decides to hold payment: the apparently more compatible part-time format already exceeds his reserve under these assumptions, and the actual start remains unconfirmed.
Resolve the agreement before making a commitment
Codesmith's regulatory-information page currently links a school catalog labeled 2024–2025. That label alone does not establish the cancellation, refund or installment rules for a future cohort. A checked terms-and-conditions route returned a 404. This review therefore does not validate a current deposit amount, interview-attempt allowance, refund percentage or payment schedule.
Beck prepares an unsent enrollment request: identify the exact program and cohort, confirmed start and live schedule; provide the governing catalog and agreement; state total charges and installment dates; explain deposit treatment, withdrawal deadlines, postponement options and what happens if the cohort does not begin as planned. He also asks which project and hiring services are included and how access ends or continues after graduation.
The current page advertises continuing hiring support for resumes, interview preparation, negotiation and alumni networking; delivery was not tested. Its lifetime label does not promise unlimited appointments or employment. Confirm the delivery arrangements rather than treating employer logos or historical salary headlines as a personal outcome estimate. This guide does not use a placement percentage or salary claim because the underlying report and denominator were not audited here.
Keep the decision tied to current documents and your own constraints. Consider the project sequence against a specific engineering gap. Keep enrollment on hold if the start remains unknown or the transition is unaffordable. Use recent LandOffer roles to check where your documented project contribution matches current responsibilities. Neither course participation nor an ambitious prototype guarantees a job.


