Mythic was founded in 2012 by University of Michigan researchers Mike Henry and Dave Fick, who set out to answer a simple question: what if a chip worked the way the human brain does? That first-principles thinking became analog compute-in-memory—an architecture that processes data where it lives instead of shuttling it back and forth across power-hungry pathways. Today, that architecture is proving its value. Mythic is the only analog compute-in-memory platform delivering 100x efficiency gains for the innovators who need AI inference at scale without the unsustainable energy cost, with technology proven across data center, automotive, robotics, and defense applications. Backed by an oversubscribed $125M raise from investors including Honda, Lockheed Martin, SoftBank, and DCVC, we're scaling toward a 1,024-chiplet architecture built for the largest LLM workloads in the data center—taking on the GPU status quo directly. We don't sell what's coming. We sell proven technology that works today, built on a fundamentally different approach to computing that we’ve been refining for over a decade. Intelligence is inevitable. How we get there is a choice. And Mythic is proving that power and efficiency were never a tradeoff to begin with.
What you'll do
Participate in design and responsible for verifying complete IP macros specifically applicable for Mythic product architecture utilizing the available NVM technologies from foundries
Involve in NVM architecture for optimum performance and area
Design and verify the new circuits to handshake with the analog computation blocks
Design, optimize and verify the NVM circuits including and not limited to high voltage generator block, x-path, y-path, test and peripheral circuits
Participate in RNM and functional model of NVM circuits for verification
Responsible for verifying the functional blocks and full chip level interfacing with NVM design
Work closely with the validation engineers to expedite the silicon evaluation process
Work closely with backend teams for the production qualification releases
What you'll bring
Ph.D. or MS degree in EE or related 5-10+ years industrial experience
Hands-on experience in NVM embedded or NVM product design and verification
Hands-on experience in NVM full chip circuit simulation setup, debugging, and verification
Hands-on experience modeling circuits using Python, Verilog-A, and System Verilog Real Number Modeling
Familiar and hands-on experience some or all memory circuit block design: decoder, high voltage switches, charge pumps, HV and LV regulator, bandgap, high performance low power sense amp, monitoring, and BIST capability
Good understanding & knowledge of semiconductor device physics, and advanced process technology
Experienced in multiple generations NVM tapeout activities, layout supervision, chip planning, silicon debugging
Excellent communication skills, ability to work well across multi-functional teams
Fluent in all major CAD tools: SPICE, Cadence, Verilog simulation software
Some knowledge of scripting such as Python
Must be a self-starter, detail-oriented, and team player
Skills Required
- Ph.D. or MS degree in Electrical Engineering or related field with 5-10+ years industrial experience
- Hands-on experience in NVM embedded or NVM product design and verification
- Hands-on experience in NVM full-chip circuit simulation setup, debugging, and verification
- Hands-on experience modeling circuits using Python, Verilog-A, and SystemVerilog Real Number Modeling
- Experience designing memory circuit blocks: decoder, high-voltage switches, charge pumps, HV/LV regulators, bandgap, sense amps, monitoring, BIST
- Good understanding of semiconductor device physics and advanced process technology
- Experience with multiple-generation NVM tapeout activities, layout supervision, chip planning, and silicon debugging
- Fluent in major CAD tools: SPICE, Cadence, Verilog simulation software
- Excellent communication skills and ability to work across multifunctional teams
- Self-starter, detail-oriented, and team player
- Some knowledge of scripting (e.g., Python)
What We Do
Mythic develops cutting-edge Analog Processing Units (APUs) utilizing unique analog compute-in-memory technology. This innovative approach aims to deliver highly energy-efficient and cost-effective AI inference solutions. By overcoming the memory bottlenecks inherent in traditional digital architectures, Mythic enables unparalleled performance-per-watt for AI workloads across diverse applications, including data centers, automotive, robotics, and defense industries.
.jpeg)
