Full-Time
Staff Embedded Software Engineer
Lightmatter
201-500 employees
Photonic computing hardware for AI acceleration
Compensation Overview
$175k - $210k/yr
Boston, MA, USA
Hybrid
Hybrid work model; some on-site required in Boston, MA per policy.
 Software Engineering (1) |
|---|
- BS and 8+ years of experience, or MS and 6+ years of experience
- Degree in Computer Science, Electrical Engineering, Robotics, or a related technical field
- Industry experience with device driver and kernel side development
- Strong understanding of hardware-software integration
- Practical knowledge of communication and control in embedded systems, peripherals, and buses
- Experience with developing commercial/enterprise-level software and firmware on Linux
- Experience with Git and CMake for building and managing software projects
- Hands-on experience designing, simulating, or deploying control of autonomous systems
- Collaborate across a multidisciplinary team including Photonic, Analog, Digital, Systems and Network experts to define control plane architecture for Lightmatter's photonic interconnect products
- Define diagnostics and monitoring architecture for photonic transceivers
- Integrate with DataCenter administration capabilities using environments such as CMIS, OpenBMC and DMTF
- Debug and troubleshoot embedded systems using tools such as GDB, JTAG, and logic analyzers
- Participate in customer engagements to gather requirements and usability criteria
- Proficiency with Python or C++
- Familiarity with debugging tools such as GDB, JTAG, and logic analyzers
- Experience developing firmware for photonic transceivers or similar systems
- Experience with environments such as CMIS, OpenBMC and DMTF
- Experience with optical communication systems and technologies
Lightmatter develops photonic computing hardware to speed up AI workloads using light-based processing. Their photonic systems run AI tasks faster and use about 90% less energy than traditional processors. They target AI and autonomous vehicle workloads and collaborate with universities to advance electro-photonic systems. Revenue comes from hardware sales, software licenses, and partnerships, with a goal of bringing scalable photonic computing to market for faster, energy-efficient AI and AV applications.
Company Size
201-500
Company Stage
Series D
Total Funding
$822M
Headquarters
Boston, Massachusetts
Founded
2017
Simplify's Take
What believers are saying
- Guide DR sampling in Q4 2026 targets hyperscalers raising rack density.
- 1.6 Tbps-per-fiber Passage chiplet with Qualcomm validates demand for denser interconnects.
- Roy Kim’s Google infrastructure background strengthens enterprise productization and deployment planning.
What critics are saying
- NVIDIA and AMD can bundle tighter platforms, squeezing Lightmatter’s standalone interconnect sales.
- Guide DR and Passage L20 remain pre-volume, leaving 2026 buyers with incumbents.
- Photonic packaging failures at ASE or foundries destroy credibility after one qualification miss.
What makes Lightmatter unique
- Lightmatter uses photonics for both chip communication and AI matrix acceleration.
- Guide DR relocates lasers into chassis, preserving zero front-panel footprint.
- Passage L200 targets 200+ Tbps per package with edgeless optical I/O.
Help us improve and share your feedback! Did you find this helpful?
Your Connections
People at Lightmatter who can refer or advise you
Benefits
Health Insurance
Dental Insurance
Vision Insurance
Life Insurance
Disability Insurance
Health Savings Account/Flexible Spending Account
Unlimited Paid Time Off
Flexible Work Hours
Remote Work Options
Paid Vacation
Paid Sick Leave
Paid Holidays
Sabbatical Leave
Hybrid Work Options
Stock Options
Company Equity
401(k) Retirement Plan
Wellness Program
Mental Health Support
Gym Membership
Phone/Internet Stipend
Home Office Stipend
Commuter Benefits
Training & Development
Conference Attendance Budget
Professional Development Budget
Family Planning Benefits
Fertility Treatment Support
Adoption Assistance
Childcare Support
Elder Care Support
Relocation Assistance
Employee Referral Bonus
Tuition Reimbursement
Professional Certification Support
Mentorship Program
Meal Benefits
Employee Discounts
Company Social Events
Growth & Insights and Company News
6 month growth
↑ 0%1 year growth
↑ 5%2 year growth
↑ 6%Lightmatter has achieved a record-breaking 1.6 Tbps throughput per fibre with its Passage Co-Packaged Optics chiplet, marking a significant milestone in AI interconnect performance. The technology, developed in collaboration with Qualcomm Technologies, uses a 16-wavelength DWDM architecture and delivers up to eight times more bandwidth per fibre than existing solutions. The silicon-proven platform combines Lightmatter's photonic engine with Qualcomm Technologies' 112G PAM4 optical SerDes chiplet, addressing interconnect bottlenecks limiting AI model scaling. The architecture reduces power consumption whilst enabling connectivity for next-generation processors and switches, with Passage L-Series photonic engines on track to deliver 100 Tbps and beyond. The record-breaking Passage L-Series photonic interconnects are now available through evaluation kits for lead customer testing.
Lightmatter has unveiled vClick Optics, the industry's first detachable fibre array unit for co-packaged optics advanced packaging. The technology enables high-volume manufacturing of 3D CPO-enabled chips by allowing manufacturers to verify optical engines before final integration, reducing yield loss risks. vClick Optics achieves insertion loss below 1.5 dB and supports high-bandwidth Dense Wavelength Division Multiplexing for 32-100Tbps optical interconnects. The technology integrates SENKO's connector solutions with Lightmatter's photonic technology and is compatible with advanced packaging flows from major foundries. Developed in partnership with ASE and SENKO Advanced Components, vClick transforms permanent fibre attachment into a plug-and-play solution. Lightmatter will showcase the technology at the Optical Fiber Communication conference in Los Angeles from 15-19 March 2026.
Lightmatter has unveiled its Passage L20 optical engine, designed to halve fibre usage in datacentres through near-package optics (NPO) rather than co-packaged optics. The chip, aimed at a 2027 launch, connects via copper to switch ASICs and breaks signals into up to 32 bidirectional fibres operating at 200 Gbps each. The L20 offers 6.4 Tbps throughput whilst consuming 30 watts. Just 16 units could replace 512 traditional 200 Gbps pluggable modules in a 102.4 Tbps switch, significantly reducing power consumption. CEO Nick Harris views NPO as a temporary solution before co-packaged optics arrive around 2028. Lightmatter also introduced vClick, a surface-attach fibre array compatible with advanced packaging technologies. The company expects to begin sampling L20 chips in late 2026.
Synopsys has partnered with Lightmatter to integrate its 224G SerDes and UCIe IP into Lightmatter's Passage 3D Co-Packaged Optics platform, targeting low-latency, high-bandwidth AI infrastructure. The collaboration reinforces Synopsys' positioning in co-packaged optics and high-speed connectivity for AI chips. However, analysts suggest the deal alone is unlikely to materially shift near-term prospects compared to larger catalysts like first-quarter 2026 results due 25 February and progress integrating Ansys following recent debt issuance. Fair value estimates from the Simply Wall St Community range widely from $376,000 to $556,000 per share. With Synopsys trading at a premium valuation, investors face execution risks around AI initiatives and Ansys integration. The company's shares have declined but may be up to 36% below fair value.
Lightmatter, a leader in photonic interconnect solutions for AI, has partnered with Global Unichip Corp. (GUC) to commercialise Passage 3D Co-Packaged Optics solutions for hyperscale AI infrastructure. The collaboration combines Lightmatter's photonic interconnect technology with GUC's ASIC design and advanced packaging expertise. The Passage platform uses silicon photonics to deliver enhanced bandwidth density and power efficiency for chip-to-chip communications, addressing connectivity bottlenecks in next-generation AI and high-performance computing workloads. By extending AI cluster scale-up capabilities across multiple racks, the solution aims to improve training time and throughput for frontier AI models. GUC will integrate Passage into its ASIC designs using advanced node chiplet and packaging workflows. The partnership targets hyperscale data centre operators seeking to overcome bandwidth and power constraints.