Summer 2026
Applied AI Engineer Intern
Posted on 6/2/2026
d-Matrix
201-500 employees
Delivers memory-integrated AI compute platforms
Compensation Overview
$30 - $60/hr
Santa Clara, CA, USA
Hybrid
Master's, PhD
 AI & Machine Learning (1) |
|---|
Get referred to d-Matrix
See people who can refer or advise you
- Pursuing a Master’s or PhD in Computer Science, Electrical Engineering, Industrial Engineering, or a related field
- Strong proficiency in Python and ML frameworks (PyTorch, TensorFlow, or scikit-learn)
- Experience with data analysis and visualization (Pandas, NumPy, Matplotlib)
- Familiarity with at least one of: time-series analysis, anomaly detection, optimization, or computer vision
- Familiarity with version control (Git) and Linux-based development workflows
- Build AI agents that diagnose why hardware tests fail—clustering failure signatures, surfacing probable root causes, and helping engineers skip weeks of manual triage
- Design LLM-powered pipelines that ingest unstructured supplier and factory reports and turn them into structured, queryable data visible to the team in real time
- Prototype intelligent document workflows that reconcile financial and procurement records, flagging discrepancies that today require hours of manual cross-checking
- Benchmark multiple LLM backends (cloud and local) across your workloads to find the right cost–quality–latency trade-offs for production deployment
- Collaborate with test, quality, and operations engineers to validate that what the models say actually matches what happens on the floor
- Prior internship or project experience in manufacturing analytics, digital twin, or process optimization
- Experience with LLMs/generative AI for structured data or knowledge extraction
- Exposure to statistical process control (SPC) or Six Sigma concepts
d-Matrix provides scalable, modular AI compute hardware and software for large datacenters, prioritizing energy efficiency and reduced data movement. Its core DIMC engine embeds compute directly into programmable memory, while a fabric of low-power chiplets delivers configurable compute resources and the accompanying software optimizes performance. This combination cuts data transfers and power use, aligning hardware design with memory-based computation for AI inference. The goal is to let large datacenters run AI workloads more efficiently at scale with customizable, modular compute platforms.
Company Size
201-500
Company Stage
Series C
Total Funding
$429M
Headquarters
Santa Clara, California
Founded
2019
Get referred to d-Matrix
See people who can refer or advise you
Simplify's Take
What believers are saying
- Corsair entered full production by June 2026 and ships to priority customers.
- Parasail deployed Corsair with Hopper and Blackwell in July 2026, validating heterogeneous inference.
- d-Matrix raised $275 million in November 2025, backing expansion and IPO preparation.
What critics are saying
- Corsair only accelerates smaller inference workloads; trillion-parameter reasoning stays NVIDIA territory.
- Two acquisitions in four months create integration drag and distract engineering through 2027.
- If Blackwell and custom ASICs close inference efficiency gaps, d-Matrix becomes a niche software bundle.
What makes d-Matrix unique
- Corsair's DIMC chiplet architecture keeps compute near memory, targeting inference latency and energy.
- August 2026 Wallaroo acquisition adds orchestration software, completing d-Matrix's silicon-to-stack inference platform.
- April 2026 GigaIO deal adds rack-scale fabric, enabling heterogeneous GPU-XPU deployments.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Hybrid Work Options
Growth & Insights and Company News
6 month growth
↑ 0%1 year growth
↓ -2%2 year growth
↑ 0%Infinity has developed agentic tools that make AI chips inference-ready within days, drastically reducing the typical months-long process. The company's autonomous agent, Ignition, automatically generates and optimises low-level compute kernels, compilers and SDKs that determine chip efficiency. In a case study with d-Matrix for its Corsair inference accelerator, Infinity reached 92% of theoretical peak performance within 10 hours of hardware access and had three frontier models running end-to-end within 10 days. The breakthrough addresses a key bottleneck in AI chip adoption: the absence of mature software stacks like Nvidia's CUDA, which took 20 years to develop. Founded in August 2025 and headquartered in San Francisco, Infinity raised $15 million in seed funding from Touring Capital and angel investors.
d-Matrix Acquires Wallaroo.ai to Accelerate AI Deployment. Second acquisition in four months brings ease of deployment and orchestration expertise to d-Matrix's silicon-to-software data center inference stack d-Matrix, the pioneer in ultra-low-latency AI inference for data centers, announced the acquisition of Wallaroo.ai, a leader in AI inference deployment and orchestration software. The acquisition brings d-Matrix the technology platform, intellectual property, and expert engineering talent of Walaroo.ai. The integration of Wallaroo follows the acquisition of GigaIO's data center business in April, and further advances d-Matrix as a category leader for rack-scale heterogeneous AI solutions that pair GPUs with specialty XPUs to achieve maximum speed and energy efficiency. With Wallaroo, d-Matrix now offers an end-to-end inference platform spanning high-performance silicon to deployment software making it easy for customers to deploy and scale low-latency inference from single-server nodes to multi-rack data center scale. The members of Wallaroo's engineering, product, and go-to-market teams joining d-Matrix bring deep expertise in software architecture, high-performance computing, Kubernetes operations, and AI inference systems. "Our customers are deploying AI inference across increasingly complex, heterogeneous environments and they've told us the biggest barrier isn't just performance, it's also the operational complexity of getting there," said Sid Sheth, founder and CEO of d-Matrix. "Wallaroo solves that. By integrating deployment and orchestration software directly into our stack, we're giving customers the simplest, fastest path from evaluation to production-scale inference. We believe AI infrastructure should expand intelligence while expanding efficiency and that starts with making it effortless to deploy." "What drew us to d-Matrix was our shared belief that inference requires seamless deployment and scale, not just faster chips," said Vid Jain, Founder and CEO of Wallaroo.AI. "Combining our AI orchestration software with d-Matrix's purpose-built silicon immediately creates an inference platform that's leaps and bounds ahead of the market. We're thrilled to join a company with this level of vision and execution, as well as with an extraordinary culture. We're very excited to build the future together."
d-Matrix has acquired Wallaroo.ai, a leader in AI inference deployment and orchestration software. This marks d-Matrix's second acquisition in four months, following the purchase of GigaIO's data center business in April. The acquisition brings Wallaroo's technology platform, intellectual property, and engineering talent to d-Matrix. It enables d-Matrix to offer an end-to-end inference platform spanning high-performance silicon to deployment software, supporting deployment from single-server nodes to multi-rack data center scale. d-Matrix's flagship Corsair inference platform is now in full production and shipping to priority customers. In early July, d-Matrix and Parasail announced the deployment of Corsair alongside NVIDIA Hopper and Blackwell GPU architectures in Parasail's cloud data center. The company is hiring engineers across several speciality areas to support its continued growth.
Parasail is deploying d-Matrix Corsair inference accelerators alongside NVIDIA Hopper and Blackwell GPUs to deliver up to 10 times faster, more cost-efficient inference services to customers. The deployment marks one of the first commercial-scale examples of heterogeneous disaggregated inference in production. The approach combines NVIDIA GPUs for compute-intensive prefill with d-Matrix Corsair accelerators for latency-sensitive decode. Parasail's automatic kernel optimisation technology dynamically routes workloads to appropriate hardware to maximise performance across its heterogeneous fleet. D-Matrix Corsair's performance stems from its Digital In-Memory Compute chiplet architecture, which integrates compute with memory on the same silicon, enabling up to 10 times faster interactive inference and up to three times better energy efficiency versus traditional approaches. The companies plan to share detailed performance results following initial deployments across Parasail's global fleet of over 40 data centres in 15 countries.
d-Matrix has announced its Corsair inference accelerator won the 2026 AI Breakthrough Award for "AI Processor Innovation". The Santa Clara-based company's platform was recognised for its ability to work alongside GPUs in heterogeneous compute environments, delivering significant increases in AI interactivity. Corsair, which entered full production this month, is purpose-built for the decode phase of disaggregated inference. The system addresses latency demands from agentic AI workloads, including real-time voice agents and interactive coding tools, whilst working with GPUs to handle compute-intensive tasks. Founded seven years ago, d-Matrix specialises in low-latency AI inference for data centres. The AI Breakthrough Awards programme, now in its ninth year, received over 5,000 nominations from companies across more than 20 countries.