Engineering Manager, Device AI

Apkudo

About The Position

The Device AI team builds the computer vision and machine learning systems that power automated cosmetic inspection, grading, and quality assessment for devices at scale. We are shipping front/back and side inspection models, advancing cosmetic inspection pipelines for partners including RSA, RXS, and MAC, and expanding into Samsung and T-Mobile for Business. We need an Engineering Manager who understands AI/ML systems deeply enough to lead technical design conversations — and who is on the keyboard enough that the team never wonders whether their manager gets it. You are a player-coach: 60% of your time is spent leading the team — on process, people, and product partnership — and 40% you are a working engineer, contributing directly to the models and pipelines your team depends on.

Requirements

BS in Computer Science or related field (relevant experience may substitute for and/or augment relevant degrees).
7+ years of professional software engineering experience, with at least 2 years in an engineering management or tech lead role.
Hands-on Python proficiency — you are comfortable writing production code and doing real code review, not just reading summaries.
Strong PostgreSQL and SQL experience; comfort with cloud infrastructure (AWS or equivalent).
Proven track record of shipping software on time with a distributed, cross-functional team.
Experience with agile delivery — sprint planning, backlog management, velocity tracking, and retrospectives.
Demonstrated ability to give direct, constructive performance feedback and support engineers' career growth.
Strong written and verbal communication skills; ability to represent engineering clearly to product, operations, and leadership audiences.
High empathy, low ego — you care more about your team succeeding than about being the smartest person in the room.
Hands-on experience with computer vision, image processing, or machine learning systems in production — you have shipped models, not just trained them.
Familiarity with ML engineering practices: model versioning, evaluation pipelines, data labeling workflows, and the experimentation-to-production lifecycle.
Experience managing teams that blend research-oriented and product-oriented engineering work — and a track record of keeping both moving.

Nice To Haves

Exposure to edge deployment, on-device inference, or hardware-in-the-loop AI systems is a meaningful plus.

Responsibilities

Own the health, velocity, and morale of your team — running effective sprints, standups, retrospectives, and one-on-ones that keep engineers growing and unblocked.
Provide structure and predictability: maintain a well-groomed backlog, own sprint commitments, and ensure the team consistently delivers against its roadmap.
Partner closely with the Product Manager on capacity planning — translating roadmap priorities into realistic sprint plans, surfacing trade-offs early, and flagging risks before they become problems.
Recruit, interview, and onboard engineers; build a team culture defined by ownership, craft, and psychological safety.
Mentor engineers at every level through regular one-on-ones, career development conversations, goal-setting, and performance feedback.
Serve as the primary escalation point for cross-team dependencies, blockers, and coordination with other pods.
Translate technical complexity into clear updates for product and leadership stakeholders.
Navigate the unique delivery dynamics of AI/ML work — managing experimentation cycles, model evaluation cadences, and the uncertainty inherent in research-adjacent engineering alongside predictable product delivery.
Spend approximately 40% of your time as an individual contributor — writing production Python code, reviewing pull requests with substantive technical feedback, and pairing with engineers on hard problems.
Lead by example in code quality, testing discipline, and documentation standards.
Contribute to architectural decisions and technical design reviews, ensuring the team's technical direction is sound and well-documented.
Contribute directly to computer vision pipelines, model evaluation infrastructure, or MLOps tooling — staying close enough to the technical work to provide meaningful guidance and unblock engineers on hard problems.