Build production AI systems for enterprise environments.
Menemsha brings enterprise GPU infrastructure into reliable model production. Our teams work across cluster platforms, training, evaluation, serving, product, security, and customer delivery.
Start with the work you are proudest of.
A title or degree can provide useful context, but it is not a substitute for evidence. Show us the systems, research, products, or customer outcomes that best demonstrate how you think and execute.
- Engineering and researchPoint to production systems, experiments, benchmarks, incident response, open source work, or technical writing.
- Product and businessDescribe product decisions, enterprise programs, customer results, or operating processes that you owned.
- In your applicationPut the most relevant example near the top. Explain your contribution, the constraints, and the result.
Engineering
7 openingsML Infrastructure Engineer
Own the compute and platform foundation that determines how reliably Menemsha can move enterprise AI workloads from accepted hardware into production.
The role
You will build the shared substrate for training, post-training, inference, and agent workloads across private GPU clusters and cloud environments. This is broad systems ownership: scheduling and isolation, secure execution, networking, state and recovery, observability, and the interfaces that let engineers use the platform without understanding every layer beneath it.
What you’ll do
- Operate CPU and GPU clusters as one coherent platform using Kubernetes, Slurm, or comparable schedulers, with clear policies for placement, quotas, autoscaling, and multi-tenant isolation.
- Design sandboxed environments for agents and untrusted code, including identity, secrets, controlled egress, artifact access, and auditable policy enforcement.
- Build checkpoint, state-management, recovery, and resumption paths for long-running training, inference, and agent workloads.
- Own connectivity across clusters and hosts: routing, peering, load balancing, service discovery, network isolation, and integration with AI fabrics.
- Automate environment creation, workload launch, telemetry, and incident diagnosis so successful field patterns become reusable platform capabilities.
Relevant experience
- Typically 4+ years building or operating distributed infrastructure, platform engineering, SRE, or HPC systems in production. Equivalent evidence of depth is also welcome.
- Hands-on ownership of Kubernetes, Slurm, Ray, or comparable schedulers and runtimes across multi-node environments.
- Experience diagnosing failures across Linux hosts, containers, storage, networking, and workload orchestration.
Basic qualifications
- Fluency in Python, Go, or another systems language, with deep Linux debugging instincts.
- Working knowledge of GPU software stacks, container runtimes, storage and data paths, networking, and platform security.
- A record of designing for failure: observable services, safe rollouts, capacity controls, recovery procedures, and clear operational ownership.
- Ability to translate researcher and customer needs into durable abstractions rather than one-off scripts.
Bonus experience
- Custom Kubernetes operators, schedulers, device plugins, CNI work, or multi-cluster control planes.
- InfiniBand or RoCE, BGP, EVPN/VXLAN, DPUs, or GPU-aware network isolation.
- Infrastructure for tool-using agents, secure code execution, or heterogeneous CPU/GPU workflows.
Thank you, there.
We received your application for ML Infrastructure Engineer. Menemsha’s hiring team will review it and contact you if there is a fit.
Distributed Training Engineer
Make large-model pretraining runs efficient, observable, and recoverable across hundreds of accelerators.
What you’ll do
- Tune data, tensor, pipeline, and expert parallel strategies for customer workloads.
- Diagnose NCCL/RCCL, straggler, memory, dataloader, and checkpoint bottlenecks.
- Establish throughput, MFU, stability, and scaling baselines.
Relevant experience
- Typically 3+ years in distributed training, machine-learning systems, or applied research engineering. Equivalent open source or research depth is also welcome.
- Hands-on experience running transformer training across multiple GPU hosts and recovering failed or degraded runs.
- Experience profiling throughput, memory use, communication, and input pipelines with clear experimental controls.
Basic qualifications
- Fluency with PyTorch distributed, Megatron, DeepSpeed, FSDP, or equivalents.
- Strong Python skills and working knowledge of Linux, containers, schedulers, checkpoints, and distributed failure modes.
- A rigorous approach to profiling, experiment design, written results, and reproducible performance baselines.
Bonus experience
- Large language model pretraining, mixture-of-experts, long-context training, or multimodal training experience.
- NCCL or RCCL tuning, InfiniBand or RoCE debugging, topology-aware placement, or network telemetry.
- Contributions to training frameworks, distributed runtimes, checkpoint systems, or performance tooling.
Thank you, there.
We received your application for Distributed Training Engineer. Menemsha’s hiring team will review it and contact you if there is a fit.
ML Post-Training Engineer
Build the systems that turn a capable base model into a reliable reasoning, tool-use, and enterprise task model.
The role
You will own post-training as a continuous distributed system: data and environment preparation, rollout generation, reward and evaluation signals, training, checkpoint movement, and release gates. The work spans researcher-facing frameworks and the production infrastructure required to run heterogeneous inference and training loops efficiently and recover from failures.
What you’ll do
- Architect post-training workflows that move cleanly from single-GPU experiments to distributed production runs across private infrastructure.
- Build and tune continuous inference, rollout, reward-model, critic, and training loops across heterogeneous GPU and CPU resources.
- Implement supervised fine-tuning and preference or reinforcement-learning methods such as DPO, GRPO, PPO, and reward modeling where they fit the customer objective.
- Integrate high-throughput inference engines and distributed runtimes, with explicit handling for placement, resharding, checkpointing, elastic recovery, stragglers, and resource contention.
- Create evaluation suites, regression gates, experiment ledgers, and human-review loops that connect model changes to measurable quality and deployment economics.
Relevant experience
- Typically 3+ years in machine-learning systems, applied research engineering, or distributed deep learning, including hands-on model training ownership.
- Experience taking at least one post-training, fine-tuning, preference-learning, or reinforcement-learning workflow beyond a notebook into a repeatable system.
- Experience running multi-GPU workloads and debugging training, inference, data, or evaluation failures end to end.
Basic qualifications
- Strong Python and PyTorch skills, including distributed training primitives and the ability to debug framework or runtime behavior below the training script.
- Practical knowledge of SFT, preference optimization, reinforcement learning, reward modeling, evaluation design, and data-quality failure modes.
- Experience with distributed systems, Kubernetes or Ray, and production failure recovery for accelerator workloads.
- Clear technical communication with both research and infrastructure teams.
Bonus experience
- Framework work in VeRL, OpenRLHF, NeMo-RL, TorchTitan, DeepSpeed-Chat, or comparable systems.
- Integration of vLLM, SGLang, TensorRT-LLM, or another serving engine into rollout generation.
- Multi-turn tool use, agent environments, visual-language post-training, or frontier-scale distributed runs.
Thank you, there.
We received your application for ML Post-Training Engineer. Menemsha’s hiring team will review it and contact you if there is a fit.
AI Systems Performance Engineer
Find and remove the bottlenecks that keep expensive GPU systems from delivering useful model throughput.
The role
You will lead end-to-end performance work across distributed pretraining, post-training, and inference. You will move across abstraction layers, from model behavior and framework overhead to kernels, memory, collectives, host networking, storage, and fabric topology. You will turn measurements into production changes and defensible capacity decisions.
What you’ll do
- Profile complete AI workloads and isolate limits across compute, memory, communication, scheduling, input pipelines, parallelism strategy, kernels, and framework runtime.
- Establish repeatable benchmarks and performance models for training throughput, MFU, rollout generation, inference latency, scaling efficiency, and cost per useful output.
- Tune GPU placement, NCCL collectives, InfiniBand or RoCE behavior, topology alignment, congestion controls, and GPUDirect paths.
- Build production-quality instrumentation and analysis tooling with PyTorch Profiler, Nsight Systems or Compute, NCCL tests, DCGM, and network telemetry.
- Translate workload evidence into changes across software, cluster configuration, operating policy, vendor escalations, and future infrastructure design.
Relevant experience
- Typically 4+ years in performance engineering, distributed ML systems, GPU computing, or HPC, with ownership of production-scale investigations.
- Experience optimizing multi-node accelerator workloads and proving improvements with repeatable measurements rather than isolated benchmark wins.
- Experience debugging across at least several layers of the stack: model, framework, kernels, drivers, hosts, collectives, network, or storage.
Basic qualifications
- Strong understanding of GPU architecture, CUDA execution, memory hierarchy, collective communication, and multi-node scaling.
- Fluency with PyTorch or JAX and practical experience with distributed parallelism, profilers, benchmarking, and experiment design.
- Ability to debug across application, framework, driver, host, switch, and fabric layers without losing sight of the production outcome.
- Strong Python skills; C++ or CUDA experience for lower-level investigation and tooling.
Bonus experience
- Large language model pretraining, RL post-training, mixture-of-experts, or high-throughput inference optimization.
- NCCL internals, RDMA, NVLink, Spectrum-X, Quantum InfiniBand, ConnectX, or BlueField experience.
- Performance modeling or simulation used to guide cluster, network, or accelerator architecture decisions.
Thank you, there.
We received your application for AI Systems Performance Engineer. Menemsha’s hiring team will review it and contact you if there is a fit.
Inference Systems Engineer
Build high-throughput, governed serving systems for enterprise models and latency-sensitive AI applications.
What you’ll do
- Benchmark and optimize model serving across latency, throughput, quality, and cost.
- Own batching, parallelism, quantization, caching, routing, and autoscaling decisions.
- Define production SLOs and failure-recovery paths.
Relevant experience
- Typically 3+ years in ML systems, backend infrastructure, or performance engineering, including production model-serving ownership.
- Experience operating latency-sensitive or high-throughput APIs under measurable reliability and capacity targets.
- Experience benchmarking generative-model inference and turning results into architecture or configuration changes.
Basic qualifications
- Production experience with vLLM, TensorRT-LLM, Triton, SGLang, or similar.
- Strong understanding of GPU memory and transformer inference behavior.
- Strong Python skills and practical knowledge of observability, load testing, deployment automation, and incident response.
Bonus experience
- Quantization, speculative decoding, custom kernels, prefix caching, or disaggregated serving experience.
- Multi-tenant GPU serving, autoscaling, admission control, routing, or workload isolation.
- Serving private or regulated AI workloads with governance, audit, or data-boundary requirements.
Thank you, there.
We received your application for Inference Systems Engineer. Menemsha’s hiring team will review it and contact you if there is a fit.
AI Infrastructure Security Engineer
Make private AI platforms defensible without making them impossible for research and engineering teams to use.
What you’ll do
- Design identity, secrets, segmentation, artifact, and data-access controls for GPU environments.
- Build audit evidence and hardening patterns for regulated customers.
- Threat-model model supply chains, training data, endpoints, and operator access.
Relevant experience
- Typically 4+ years in cloud, platform, product, or infrastructure security, including direct ownership of production controls.
- Experience securing Kubernetes or Linux platforms in enterprise or regulated environments.
- Experience leading threat models, security reviews, remediation work, and evidence collection with engineering teams.
Basic qualifications
- Fluency with IAM, Kubernetes security, Linux hardening, and policy-as-code.
- Working knowledge of secrets management, network segmentation, software supply-chain controls, logging, and incident response.
- Ability to translate controls into practical engineering workflows and explain tradeoffs to technical and executive stakeholders.
Bonus experience
- Security work in AI infrastructure, GPU clusters, HPC, model serving, or agent execution environments.
- Experience with SOC 2, ISO 27001, NIST, FedRAMP, HIPAA, PCI, or comparable assurance programs.
- Container sandboxing, confidential computing, zero-trust networking, artifact signing, or model supply-chain security.
Thank you, there.
We received your application for AI Infrastructure Security Engineer. Menemsha’s hiring team will review it and contact you if there is a fit.
Forward-Deployed AI Engineer
Lead technical delivery inside customer environments, from first cluster diagnostic through production handoff.
What you’ll do
- Own deployments across infrastructure, model workflow, and customer operating process.
- Debug live systems, communicate tradeoffs, and keep executive and technical stakeholders aligned.
- Turn successful engagements into reusable product and playbooks.
Relevant experience
- Typically 5+ years across ML infrastructure, platform engineering, solutions architecture, or technical delivery.
- Experience leading ambiguous customer deployments from discovery through production handoff and measurable operating results.
- Experience debugging live systems while coordinating customer engineers, vendors, and business stakeholders.
Basic qualifications
- Broad ML systems depth and strong customer-facing instincts.
- Strong Python and Linux skills with working knowledge of GPU compute, networking, storage, schedulers, and model workflows.
- Excellent written communication, structured program judgment, and a low-ego operating style.
- Ability and willingness to travel up to 30% for customer work.
Bonus experience
- Private AI, on-prem GPU, HPC, distributed training, or high-throughput inference deployment experience.
- Work with regulated enterprises, air-gapped environments, or complex security and data-boundary requirements.
- A record of turning field solutions into reusable software, reference architectures, or delivery playbooks.
Thank you, there.
We received your application for Forward-Deployed AI Engineer. Menemsha’s hiring team will review it and contact you if there is a fit.
Go-to-market & operations
3 openingsEnterprise Account Executive
Build a focused book of enterprise customers investing in private GPU infrastructure and accountable model outcomes.
What you’ll do
- Develop and close complex, technical opportunities with infrastructure and AI leaders.
- Run disciplined account plans across long enterprise buying cycles.
- Partner with engineers to scope credible outcomes, not generic transformation projects.
Relevant experience
- Typically 5+ years selling enterprise infrastructure, platform software, AI systems, or high-value technical services.
- A record of originating and closing complex six- or seven-figure opportunities with multiple technical and executive stakeholders.
- Experience managing long buying cycles, procurement, security review, business cases, and expansion planning.
Basic qualifications
- Executive presence with enough technical fluency to earn engineering trust.
- Strong discovery, qualification, account planning, forecasting, negotiation, and written follow-through.
- Ability to sell measurable operating outcomes while maintaining credibility about technical limits and delivery risk.
- Ability and willingness to travel for strategic customer meetings.
Bonus experience
- Experience selling GPU infrastructure, HPC, private cloud, data platforms, model tooling, or enterprise AI services.
- Relationships with CIO, CTO, infrastructure, research, security, or AI platform leaders in regulated industries.
- Early-stage GTM experience building territory strategy, messaging, sales process, and reference customers.
Thank you, there.
We received your application for Enterprise Account Executive. Menemsha’s hiring team will review it and contact you if there is a fit.
Solutions Architect, AI Infrastructure
Translate customer constraints into a technically sound path from idle infrastructure to a production workload.
What you’ll do
- Lead technical discovery, architecture sessions, demos, and pre-engagement diagnostics.
- Develop reference designs, statements of work, and outcome-based proposals.
- Feed recurring customer needs back into product and delivery.
Relevant experience
- Typically 4+ years in solutions architecture, technical pre-sales, consulting, or customer-facing platform engineering.
- Experience translating ambiguous enterprise requirements into architectures, phased delivery plans, and credible success criteria.
- Experience presenting to both hands-on engineers and senior decision-makers through complex technical buying cycles.
Basic qualifications
- Ability to reason across compute, network, storage, ML frameworks, and security.
- Clear presentation, whiteboarding, technical writing, and requirements-management skills.
- Enough hands-on Linux and cloud or data-center depth to validate assumptions and earn engineering trust.
- Ability and willingness to travel for discovery, workshops, and customer launches.
Bonus experience
- Distributed GPU, HPC, Kubernetes, Slurm, model training, inference, or AI platform architecture experience.
- Security and compliance design for regulated, private, hybrid, or air-gapped environments.
- Experience writing statements of work, reference architectures, capacity models, or outcome-based proposals.
Thank you, there.
We received your application for Solutions Architect, AI Infrastructure. Menemsha’s hiring team will review it and contact you if there is a fit.
Customer Success & Delivery Lead
Own the operating cadence that keeps multiple technical engagements moving toward measurable customer outcomes.
What you’ll do
- Run delivery plans, risk reviews, stakeholder updates, and handoffs across 10 to 12 strategic clients.
- Track adoption and outcome evidence after technical milestones land.
- Build the lightweight systems that let a senior team scale without bureaucracy.
Relevant experience
- Typically 5+ years in enterprise technical delivery, customer success, implementation leadership, or program management.
- Experience owning a portfolio of strategic accounts or concurrent technical programs with executive visibility.
- Experience guiding customers from implementation through adoption, operating handoff, renewal, or expansion.
Basic qualifications
- Strong program judgment, executive communication, risk management, and attention to detail.
- Comfort operating around infrastructure and ML engineering teams without losing sight of customer outcomes.
- Ability to build clear delivery cadences, decision logs, outcome evidence, and escalation paths.
- Ability and willingness to travel for key customer milestones.
Bonus experience
- Delivery experience in AI infrastructure, cloud platforms, data systems, cybersecurity, or enterprise software.
- Experience serving regulated customers or coordinating across vendors, integrators, and internal engineering teams.
- Early-stage operations experience creating scalable systems before dedicated tooling or large support teams exist.
Thank you, there.
We received your application for Customer Success & Delivery Lead. Menemsha’s hiring team will review it and contact you if there is a fit.
Research & evaluation
2 openingsApplied AI Research Engineer
Develop and validate methods that improve model quality for enterprise workloads and operating environments.
What you’ll do
- Design experiments for domain adaptation, post-training, retrieval, and agent workflows tied to defined customer objectives.
- Build datasets, baselines, evaluation suites, and analysis tools that make model changes measurable.
- Work with infrastructure and delivery engineers to move successful experiments into repeatable production workflows.
Relevant experience
- Typically 3+ years in applied machine learning, research engineering, or model development.
- Experience running controlled experiments on transformer models and interpreting quality, latency, and cost tradeoffs.
- Experience translating an open research question into a scoped technical plan and a clear written conclusion.
Basic qualifications
- Strong Python and PyTorch skills with practical knowledge of training, inference, data preparation, and evaluation.
- Sound experimental design, statistical judgment, and attention to reproducibility.
- Ability to communicate methods and results to research, engineering, and customer teams.
Bonus experience
- Supervised fine-tuning, preference optimization, reinforcement learning, agent evaluation, or multimodal model experience.
- Research or delivery work involving private data, regulated environments, or constrained infrastructure.
- Publications, technical writing, open source contributions, or other evidence of independent technical work.
Thank you, there.
We received your application for Applied AI Research Engineer. Menemsha’s hiring team will review it and contact you if there is a fit.
Model Evaluation Engineer
Build the measurement systems that determine whether a model is ready for a production task.
What you’ll do
- Create task-specific evaluation suites, regression tests, scorecards, and release criteria for enterprise models.
- Develop data pipelines and human review workflows that produce reliable evaluation evidence.
- Investigate quality regressions and connect model behavior to training data, prompts, serving configuration, and operating context.
Relevant experience
- Typically 3+ years in model evaluation, applied machine learning, data science, ML quality, or a related discipline.
- Experience designing evaluations for language, reasoning, retrieval, agent, or multimodal systems.
- Experience using evaluation results to support model, product, or release decisions.
Basic qualifications
- Strong Python skills and practical knowledge of statistics, dataset construction, sampling, and error analysis.
- Ability to distinguish measurement noise, data leakage, evaluator bias, and genuine model change.
- Clear written communication and a disciplined approach to versioning data, prompts, metrics, and results.
Bonus experience
- Experience with red teaming, safety evaluations, LLM judges, human labeling systems, or production monitoring.
- Evaluation work in healthcare, financial services, government, legal, or another regulated domain.
- Experience connecting offline evaluations with online reliability, latency, adoption, or business outcomes.
Thank you, there.
We received your application for Model Evaluation Engineer. Menemsha’s hiring team will review it and contact you if there is a fit.
Product
2 openingsProduct Manager, AI Infrastructure
Set product direction for the tools and operating workflows that move private GPU environments into production use.
What you’ll do
- Define product priorities across cluster diagnostics, workload launch, observability, evaluation, serving, and operating handoff.
- Conduct customer discovery and convert repeated field needs into clear requirements, roadmaps, and release decisions.
- Partner with engineering and delivery teams to measure adoption, reliability, customer outcomes, and product gaps.
Relevant experience
- Typically 5+ years in technical product management for infrastructure, developer tools, data platforms, or machine learning systems.
- Experience taking an enterprise product from customer discovery through launch, adoption, and iteration.
- Experience working directly with technical buyers and users in complex operating environments.
Basic qualifications
- Technical fluency across compute, networking, storage, cloud or data center platforms, and software operations.
- Strong product judgment, requirements writing, prioritization, and quantitative decision making.
- Ability to work closely with engineers and customers while maintaining a clear product strategy.
Bonus experience
- GPU infrastructure, distributed systems, model training, inference, or enterprise AI platform experience.
- Early-stage product experience establishing roadmaps, operating metrics, and customer feedback systems.
- Experience with enterprise pricing, packaging, security review, or deployment across private environments.
Thank you, there.
We received your application for Product Manager, AI Infrastructure. Menemsha’s hiring team will review it and contact you if there is a fit.
Product Designer, Technical Systems
Design clear workflows for engineers and operators managing complex AI infrastructure and model production systems.
What you’ll do
- Design workflows for diagnostics, workload configuration, telemetry, evaluation review, incident response, and operating handoff.
- Conduct research with platform engineers, ML practitioners, operators, and enterprise stakeholders.
- Build prototypes and design systems that support a consistent product across technical surfaces.
Relevant experience
- Typically 4+ years designing complex business or technical software products.
- Experience simplifying dense workflows, configuration, data, and system state for expert users.
- Experience working through the full product cycle from research and prototyping to implementation and iteration.
Basic qualifications
- A portfolio demonstrating strong interaction design, information architecture, visual hierarchy, and systems thinking.
- Proficiency with modern design and prototyping tools and the ability to work closely with frontend engineers.
- Strong user research, written communication, and product reasoning skills.
Bonus experience
- Experience with developer tools, cloud consoles, observability products, data platforms, or security software.
- Data visualization experience for performance, reliability, capacity, or model quality workflows.
- Early-stage product experience establishing design practices and contributing directly to product strategy.
Thank you, there.
We received your application for Product Designer, Technical Systems. Menemsha’s hiring team will review it and contact you if there is a fit.
A clear interview process for every role.
We evaluate demonstrated ability, judgment, and relevant experience. Production work, customer results, research, open source contributions, and technical writing can all provide useful evidence.
- 01Application review
We review your experience and work samples against the responsibilities and qualifications listed for the role.
- 02Introductory interview
A video conversation with the hiring manager or a team member about your experience, interests, and the role. You will have time to ask questions.
- 03Role interviews
Technical roles include live problem solving, coding, or system design. Business and product roles use structured conversations and relevant scenarios. We do not assign unpaid projects outside the interview process.
- 04Final conversations and decision
You will meet additional team members, discuss how the group works, and ask any remaining questions. We may conduct references before a final decision.
Your background may fit more than one role.
Apply to the position that best matches your experience. We will consider relevant opportunities on adjacent teams during the review.