A new engineer joining a Vietnam-based dedicated team should have a merged pull request in a live environment inside the first two weeks, own a bounded module without a reviewer rewriting it by day 45, and be trusted to make one architectural call unsupervised by day 90. If none of those three things happen on schedule, the problem is rarely talent. It is almost always onboarding design, unclear ownership, or a missing checkpoint that nobody defined before the engineer started. Judging a hire before day 90 on anything other than these shippable milestones tells you more about the client's ramp process than the engineer's ability.
TL;DR
Time-to-productivity for offshore engineers typically runs 1 to 2 months for mid-level roles and 3 to 6 months for senior roles on complex codebases. Judging output before that window closes measures onboarding, not talent.
The first 90 days should be structured around three checkpoints: first merged commit (week 1-2), first owned module (day 30-45), first independent architectural decision (day 75-90).
Junior and senior engineers in Vietnam ship at different speeds by design. Junior engineers ($900-$1,300/month) execute defined tasks; senior engineers ($2,800-$4,500+/month) own modules and mentor, which is why the milestones differ by seniority.
A dedicated team model with standardized onboarding, an assigned buddy, and early shippable work outperforms an ad hoc "figure it out" ramp, both in engagement quality and time-to-productivity.
AI coding tools, specifically Claude Code, let a new engineer's first commit be a working feature rather than a config file, because the tool handles boilerplate while the engineer learns the codebase.
About the Author: 724SOFTWARE has built and scaled dedicated engineering teams for clients in fintech, healthcare, and SaaS across 10+ countries, with 200+ engineers and a 95% client retention rate. This piece draws on the onboarding checkpoints the company uses when embedding new engineers into client workflows under its ODC and staff augmentation models.
What Does "First 90 Days" Actually Mean for an Engineering Hire?
The first 90 days is the window during which a new engineer moves from reading documentation to owning production code without supervision. It is not a probation period in the HR sense. It is a ramp curve, and the curve has a documented shape: engineers in offshore or distributed teams typically reach baseline productivity in 1 to 2 months, with senior roles or complex codebases extending that to 3 to 6 months. Anyone benchmarking a new hire against day-30 output alone is measuring against a number the industry itself doesn't expect to be met yet.
The mistake many hiring managers make is treating the first 90 days as a single evaluation point rather than three separate checkpoints, each testing a different thing. Building the operating system the team runs on, meaning what "done" looks like, where decisions get made, and how priorities are communicated, is itself a 90-day project for new engineering leadership. The same logic applies one level down: a new engineer needs that operating system defined for them before their output can be judged fairly.
What Should Ship in Week One?
Week one should produce one merged, deployed change, however small. This is the first-commit milestone, and it exists to test the onboarding pipeline, not the engineer.
A new engineer's first two weeks typically involve environment setup, codebase orientation, and shadowing existing pull requests. Structured onboarding practices at companies that do this well assign a buddy, follow a 30-60-90 day plan, and deliberately front-load small, shippable work rather than reading tasks, specifically to get a first commit into production fast. Getting an early contribution shipped, alongside steady relationship-building and hitting agreed milestones, is one of the ways new engineers are measured during onboarding.
Here is what a realistic week-one deliverable looks like on a well-run dedicated team:
Local environment fully configured and able to run the existing test suite
One code review completed on someone else's PR, to build a mental model of team conventions
One small fix or feature merged and deployed, ideally something with a visible outcome (a bug fix, a UI tweak, a config change)
A documented list of open questions about the codebase, shared with the assigned buddy
If week one produces none of this, check the onboarding checklist before checking the engineer's resume.
What Should Ship by Day 45?
By day 45, a new engineer should own one bounded module end to end, meaning they can take a ticket in that module from requirement to deployed fix without a senior engineer rewriting the core logic. This is the point at which the engineer stops shadowing and starts contributing independently on a defined slice of the system.
This checkpoint maps directly to the documented productivity gap between junior and senior engineers in Vietnam's tech sector. Junior engineers are expected to execute well-defined tasks with clear instructions; senior engineers are expected to own complex modules, make architectural decisions, and mentor others. The market reflects this: junior developers in Vietnam typically earn $900 to $1,300 per month, while senior engineers command $2,800 to $4,500 or more, a gap driven by faster delivery and independent problem-solving. The day-45 checkpoint should be calibrated to seniority: a junior engineer owning one module cleanly is a strong result; a senior engineer should be doing this within the first two to three weeks.
Stepping back from the individual milestone, the harder question for the hiring team is whether the module boundary was actually clear enough to hand off. A vague module ("the reporting stuff") produces a vague 45-day result no matter how capable the engineer is. Defining ownership boundaries in week one is a prerequisite for a fair day-45 judgment.
What Should Ship by Day 90?
By day 90, a new engineer should have made at least one architectural or design decision without it being reversed by review. This does not mean redesigning a system. It means being trusted to choose between two reasonable implementation approaches and defending that choice, unsupervised, in a code review.
This is the point where a new engineering hire transitions from "executing the plan" to "contributing to the plan," and it is the clearest, most defensible signal of hire quality because it tests judgment rather than typing speed. New hire satisfaction and steady, dependable delivery are the kinds of signals used alongside this milestone, because an engineer who has reached this point but is unhappy with the role or team is a retention risk regardless of technical output.
A related but distinct question at this stage is whether the engineer has started mentoring or reviewing others. That expectation belongs to senior hires specifically, not juniors at day 90. Conflating the two produces unfair scorecards on both sides.
How Are Claude Code Developers Changing the Shape of the First 90 Days?
A Claude Code developer is an engineer who uses Anthropic's Claude Code as a standard part of their daily workflow, not as an experiment bolted onto existing habits. This changes the ramp curve in one specific way: the week-one first commit is more likely to be a working feature than a scaffolding change, because the tool reduces the time spent on boilerplate and lets the engineer spend more of week one reading and understanding the existing codebase instead of writing repetitive setup code.
724SOFTWARE is a selected Anthropic partner in Vietnam and trains its engineers to use Claude Code as part of normal delivery work. In practice, this means a new engineer joining a client's dedicated team already has the habit of using Claude Code to explore an unfamiliar codebase (asking it to summarize a module before touching it) rather than reading files cold. That does not change the 90-day judgment framework described above. It changes how fast an engineer clears the early orientation work that used to eat into week one and week two.
What Should a Client Do If a New Hire Isn't Shipping on Schedule?
The first step is not to replace the engineer. It is to check which of the three checkpoints failed and why. A missed week-one commit usually points to environment setup friction or an undocumented codebase. A missed day-45 module ownership usually points to unclear scope or a bottlenecked reviewer. A missed day-90 architectural decision usually points to a culture where every decision still routes through one senior person, regardless of who is new.
Frequently Asked Questions
How long should it take a new offshore engineer to become productive?
1 to 2 months for mid-level roles, extending to 3 to 6 months for senior roles or highly complex codebases, based on industry benchmarks for distributed and offshore teams.
What is the difference between a junior and senior engineer's first 90 days?
A junior engineer's 90 days should end with one owned module and clean execution on defined tasks. A senior engineer's 90 days should end with at least one independent architectural decision and early mentoring involvement.
What metrics should I track during a new engineer's first 90 days?
Time-to-first-commit, time-to-productivity, 90-day retention rate, and new hire satisfaction score are the standard set used by structured onboarding programs.
Does using Claude Code change how fast a new engineer ramps up?
It can shorten the orientation phase within week one and two by reducing time spent on boilerplate, but it does not change the 45-day and 90-day ownership and judgment checkpoints, which depend on codebase familiarity and trust built over time.
What is a reasonable first deliverable for a new engineer's first week?
One small, merged, and deployed change, plus one completed code review on an existing pull request, is a realistic and testable week-one outcome.
Should I judge a new hire's fit before day 90?
Only on process signals, such as communication clarity and engagement in reviews, not on technical output volume, since the documented productivity ramp for offshore roles runs past the 60-day mark for most levels.
What causes a missed 90-day checkpoint more often: the engineer or the process?
In dedicated team setups, a missed checkpoint most often traces back to unclear module ownership or an undocumented onboarding path, not individual capability.
About 724SOFTWARE
724SOFTWARE is a Vietnam-based technology partner delivering dedicated teams, offshore development centers, and embedded staff augmentation for clients in fintech, healthcare, SaaS, and enterprise software. With 200+ engineers, 58% at senior level, and delivery experience across 10+ countries, the company runs standardized onboarding processes designed around the ramp checkpoints described above. As a selected Anthropic partner in Vietnam, 724SOFTWARE trains its engineers to use Claude Code as part of daily delivery, and operates under ISO 9001, ISO 27001:2022,, and GDPR compliance with follow-the-sun support across timezones and under-10-minute incident response.
If you're evaluating how a new engineer or dedicated team should ramp inside your workflow, get in touch with 724SOFTWARE at https://724software.com.vn/.
