⚑ Flagged: ghost-job signals. Risk level: high. How we flag ghost jobs →
Why this grade
This listing scored 13/100, which is an F. It lost the most ground on pay transparency.
- Description depth 20 / 20 How much the posting actually says about the work, measured in characters of real text.
- Corroboration 10 / 10 Whether more than one source carries this listing.
- Remote clarity 8 / 15 Whether "remote" means anywhere, or is quietly restricted to one country.
- Role specificity 0 / 10 Whether the listing is tagged well enough to tell what the role actually is.
- Freshness 0 / 15 How recently it was posted. Older postings are likelier to be filled or abandoned.
- Pay transparency 0 / 25 A published salary range, worth more than any other single factor because it is what a candidate cannot find out without applying.
-25 Ghost-job penalty — Deducted for signals that this posting may not be a real, currently-open role — staleness, repeated relisting, or talent-pool language.
Every figure above is arithmetic over the posting itself — its salary field, its text, its age, its tags and how many sources carry it. How the grades work →
About A1
There are over 5 billion users using basic applications today such email, notes, tasks that are not AI-native. Our mission is to build a proactive smart assistant for everyday users to bring intelligence to conversations, errands, organising and workflows, with minimal prompting.
Our product focuses on achieving high reliability for long-running workflows, persistent context, and real-world task completion. The system must handle multi-step reasoning, interact with external tools, and remain reliable despite non-deterministic model behavior. Our objective is to help users complete tasks daily enjoyable with over ~90%* reduced time.
Role
You will be responsible for turning research direction into working, production-grade ML systems. This role owns the execution layer of A1’s intelligence – training pipelines, inference systems, evaluation tooling, and deployment.
Focus
Build and own end-to-end ML pipelines spanning data, training, evaluation, inference, and deployment.
Fine-tune and adapt models using state-of-the-art methods such as LoRA, QLoRA, SFT, DPO, and distillation.
Architect and operate scalable inference systems, balancing latency, cost, and reliability.
Design and maintain data systems for high-quality synthetic and real-world training data.
Implement evaluation pipelines covering performance, robustness, safety, and bias, in partnership with research leadership.
Own production deployment, including GPU optimization, memory efficiency, latency reduction, and scaling policies.
Collaborate closely with application engineering to integrate ML systems cleanly into backend, mobile, and desktop products.
Make pragmatic trade-offs and ship improvements quickly, learning from real usage.
Work under real production constraints: latency, cost, reliability, and safety
Requirements
Strong background in deep learning and transformer-based architectures.
Hands-on experience training, fine-tuning, or deploying large-scale ML models in production.
Proficiency with at least one modern ML framework (e.g. PyTorch, JAX), and ability to learn others quickly.
Experience with distributed training and inference frameworks (e.g. DeepSpeed, FSDP, Megatron, ZeRO, Ray).
Strong software engineering fundamentals – you write robust, maintainable, production-grade systems.
Experience with GPU optimization, including memory efficiency, quantization, and mixed precision.
Comfort owning ambiguous, zero-to-one ML systems end-to-end.
A bias toward shipping, learning fast, and improving systems through iteration.
Ideal Experience
Experience with LLM inference frameworks such as vLLM, TensorRT-LLM, or FasterTransformer.
Contributions to open-source ML or systems libraries.
Background in scientific computing, compilers, or GPU kernels.
Experience with RLHF pipelines (PPO, DPO, ORPO).
Experience training or deploying multimodal or diffusion models.
Experience with large-scale data processing (Apache Arrow, Spark, Ray).
How We Work
The best products today in the world were built by small, world class teams. We are a high talent density and hands-on team. We make decisions collectively, move at rapid speed, striking a balance between shipping high quality work and learning. Joining our team requires the ability to bring structure, exercise judgment, and execute independently. Our goal is to put in hands of our users a truly magical product
Interview process
If there appears to be a fit, we'll reach to schedule 3, but no more than 4 interviews.
Applications are evaluated by our technical team members. Interviews will be conducted via virtual meetings and/or onsite.
We value transparency and efficiency, so expect a prompt decision. If you've demonstrated the exceptional skills and mindset we're looking for, we'll extend an offer to join us. This isn't just a job offer; it's an invitation to be part of a team that's bringing AI to have practical benefits to billions globally.
Originally posted on Himalayas
Apply for this role Opens himalayas.app — the link as listed; we have not yet verified it is the employer's own page
Where this listing came from
- 15 Jul 2026 Himalayas first sighting
Seen on 1 board over 30 days.