Senior Machine Learning Engineer 5, Foundational Research in Agentic AI at Adobe (Bangalore or Noida)

Application ends: October 15, 2026
Apply Now

Job Description

Adobe is hiring a Senior Machine Learning Engineer 5 to lead foundational agentic AI research in Bangalore, with Noida also listed as a location. The work is on the models and learning methods that let AI agents reason, plan, remember, use tools and cope with long, multi-step tasks. Adobe says plainly that this is not a prompt-engineering or agent-orchestration job: it wants a scientist who can invent new modelling and learning approaches rather than mainly assemble existing LLMs, prompts and agent frameworks.

At a glance

  • Position: Senior Machine Learning Engineer 5 (requisition R171066)
  • Where: Bangalore, India (Noida also listed)
  • Focus: agentic AI, foundation models, reasoning, planning, memory and tool use
  • Experience: 9+ years of hands-on ML engineering
  • Posted: 18 August 2026
  • Apply: open until filled (re-checked 15 October 2026)

What the agentic AI research role at Adobe Bangalore involves

  • Carry out fundamental and applied research on agentic AI and foundation models: reasoning, planning, memory, tool use, multimodal intelligence and long-horizon interaction
  • Design new model architectures, learning algorithms, training methods and inference techniques that make agents more capable
  • Study open problems such as grounding, adaptation, self-improvement and learning from interaction
  • Build research prototypes, design rigorous experiments, and carry validated ideas through to product
  • File intellectual property and publish at top-tier conferences
  • Lead technical design reviews, write engineering RFCs, set quality standards, and mentor junior and mid-level ML engineers

Why this role stands out

Many industry “agent” jobs are about wiring existing models together. This posting instead asks for new algorithms and architectures, with patents and publications as part of the job, which puts it closer to an industrial research scientist post than to a typical ML engineering role. It suits a researcher with a strong publication record who also wants to see their work ship in widely used products.

What Adobe is looking for

  • At least 9 years of hands-on machine learning engineering experience in industry or research
  • An MS or PhD in computer science, machine learning, artificial intelligence or a related field
  • Strong fundamentals in machine learning, deep learning, optimisation and statistical modelling
  • Research experience in at least one of: large language models and NLP; generative modelling (diffusion, flow matching, autoregressive models, VAEs); computer vision; multimodal foundation models; representation learning; reinforcement learning; reasoning and planning
  • A proven ability to frame new research problems, develop new approaches and test them experimentally, shown by a strong publication record or measurable research impact
  • Strong programming and experimentation skills with modern deep learning frameworks, and clear communication with cross-functional teams

Pay and location

The posting does not state a salary. The role is full time and based in Bangalore, with Noida listed as an additional location.

How to apply

Apply through the official Adobe job posting. Adobe’s posting does not give a closing date, so the role is open until filled; ResearchJobs.in will re-check this listing on 15 October 2026. Adobe notes that AI or recording tools may not be used in live interviews unless the interviewer invites them or they are agreed in advance as an accommodation. Check the posting for location and work-authorisation details before you apply.

See all our artificial intelligence jobs, or browse more research jobs on ResearchJobs.in.

Hiring institution: Adobe

Official advertisement: adobe.wd5.myworkdayjobs.com

How to prepare for this application

  • Lead with your research: pick one or two contributions (papers, patents or shipped models) and explain the problem, your new idea, and how you showed it worked.
  • Think below the prompt: be ready to discuss how you would train agents for planning, memory or tool use, for example with reinforcement learning from interaction or post-training, rather than how you would prompt them.
  • Revise generative modelling: the posting lists diffusion, flow matching, autoregressive models and VAEs, so expect questions on their trade-offs.
  • Evaluation design: sketch how you would test a long-horizon agent reliably, with baselines, ablations and failure analysis.
  • Seniority: prepare examples of leading design reviews, writing technical proposals and mentoring engineers.

About Adobe

Adobe is a US software company known for Photoshop, Acrobat and Creative Cloud, and for its Firefly generative AI models. It has large engineering and research teams in India, with offices in Bangalore and Noida.

We send one confirmation email first. Every alert has an unsubscribe link.