Research Scientist III, Last Mile Science
Amazon has advertised Research Scientist III, Last Mile Science in Bengaluru, Karnataka, under job ID 10586334. The listed entity is ADCI – Karnataka. Amazon Science's original careers record shows a posting date of 22 October 2026, and the official job page provides an active Apply now route.
This is an experienced research-science job, not an internship or fresher opening. The basic qualifications include a PhD or master's degree and three or more years of building machine-learning models for business applications. Salary, fixed tenure, deadline and remote work are not advertised.
Opening at a glance
| Employer | Amazon |
|---|---|
| Type | Job |
| Location | Bengaluru, Karnataka, India; remote work not advertised |
| Pay / stipend | Not disclosed |
| Duration / tenure | Not disclosed |
| Apply by | Not disclosed; apply while the listing remains open |
| Reference | 10586334 |
| Original notice / posting date | 2026-10-22 |
Research applied to delivery operations
The role sits within an analytics context focused on Amazon's delivery network. It combines research-science capability with work on practical business problems, scalable processes and operational decisions. The description emphasises improving delivery experiences through data and technology.
Candidates would take ownership of important analytical problems, assess advanced solutions and work with ambiguity. Programme-management skills and the ability to turn analysis into useful execution matter alongside technical research. This is not simply a role for producing a model without considering how its result affects operations.
The source does not specify a particular public dataset, model architecture or tool-only project that every successful candidate will own. Those details should be clarified through the team's recruitment discussions rather than invented from the title.
Basic qualifications
The formal Basic Qualifications section requires a PhD or master's degree, experience investigating how scientific principles can be applied to business problems and products, and three-plus years building machine-learning models for business applications.
Knowledge of programming languages such as C/C++, Python, Java or Perl and programming experience in a relevant language are listed. Experience with machine learning and neural deep-learning methods is also required. General interest in AI alone does not replace the stated professional and academic conditions.
Applicants should be clear about the business problem, their individual contribution, the model's evaluation and what happened when it was used. A project list without evidence of this experience should not be treated as automatically meeting a three-year threshold.
Preferred research background
The listing's preferred criteria include a quantitative PhD, substantial applied research, strong mathematics and statistics, algorithm design and product development, and publications or patents. Experience mentoring scientists, strategic judgement and communication with technical and business stakeholders are also relevant.
The source contains two preferred-experience descriptions: the narrative mentions five-plus years of extensive research, while the dedicated Preferred Qualifications section mentions three-plus years of applied research. Both are preferred descriptions; the clear basic threshold remains three-plus years building ML models for business applications together with the stated degree and other basics. Applicants should read both sections and clarify their fit with the recruiter.
Tools and frameworks mentioned include R, scikit-learn, Spark MLlib, MXNet and TensorFlow, alongside experience with real datasets and distributed systems. They indicate the team's technical context, not a promise that one certificate or a single tool guarantees selection.
Location and employment details
The role is listed in Bengaluru, India. Remote or hybrid work is not advertised, so Jobsii is not presenting it as WFH. Confirm office expectations, team arrangements and any travel or relocation considerations directly with Amazon.
The verified source does not disclose salary, compensation period, a fixed employment tenure or a closing date. Do not rely on compensation from a different Amazon level or an aggregator's estimate as though it were the employer's offer. Actual terms belong in the formal recruitment and offer process.
Preparing a relevant application
As optional preparation, select an example in which a scientific method addressed a genuine product or operational problem. Explain the assumptions, data quality, evaluation method and limitations, and how you communicated a recommendation or production result.
If you have publications, patents or distributed-system experience, describe your own contribution accurately. Remove confidential company or customer information from examples. These suggestions are editorial preparation, not a new employer test or a guarantee of interview topics.
Useful questions include how the team measures success, how research is evaluated against operational constraints, the balance between modelling and programme execution, and what mentorship responsibilities apply. Ask about undisclosed compensation and location terms before making commitments.
Apply through Amazon
Use the official Amazon Jobs page linked below, confirm the title and job ID 10586334, and follow the Apply now route. Submit truthful qualifications and work history. A vacancy page or application acknowledgement is not an offer.
Jobsii does not guarantee selection and does not collect payment for an Amazon job. Recheck the current original listing before applying and use only the employer's recruitment channels for sensitive application details.
