Last November, San Jose, CA-based embedded systems developer Zeidman Technologies filed a lawsuit against the U.S. government in the U.S. Court of Federal Claims. At issue in the case are the government’s grant award and contracting practices after multiple Small Business Innovation Research (SBIR) proposals filed by Zeidman were rejected in what Zeidman alleges was an unlawful abuse of discretion. Bob Zeidman, the company’s founder, has stated that his intent in bringing this lawsuit is to fix or replace another government bureaucracy that is wasting taxpayer money.
Zeidman Technologies’ allegations stem from an SBIR solicitation for a U.S. Department of Defense project which was open in early 2016. Titled Integrated Code Base and High Performance Embedded Computing Tool, the funding topic solicited proposals for technologies which could be used by the Air Force and the space community to serve as a tool for optimizing software code to run on a modern space processor and that could translate high level algorithm descriptions into high performance and hardware efficient implementations. The solicitation notice described current high performance embedded computing application processing as inefficient, requiring iterative rounds of software performance optimization to attain required application latency and throughput performance.
Bob Zeidman, the founder and CEO of Zeidman Technologies, has worked for more than three decades in the development of reduced instruction set computer (RISC)-based parallel processor systems, hardware emulators, application-specific integrated circuits (ASICs) and other computing architectures for real-time systems. A pioneer in software synthesis and analysis, Zeidman holds 22 patents and has written textbooks in the field. In a telephone interview, Zeidman said that a friend of his who regularly monitors SBIR notices pointed out the Department of Defense solicitation which appeared to be looking for a software systems much like those which Zeidman has developed himself.
In the lawsuit filed in the Court of Federal Claims, Zeidman Technologies alleges that the U.S. government gave three conflicting and contradictory explanations as to why Zeidman’s proposal for the SBIR request was denied, despite the fact that Zeidman owns patented technology that specifically met the requirements of the request. These rejection explanations also allegedly relied on criteria that were not explicitly stated in the SBIR request for proposals, resulting in what Zeidman argues was an arbitrary decision.
Zeidman Technologies first received an explanation as to why its proposal wasn’t approved back in August 2016. In that explanation, Zeidman was informed that the funding topic was seeking a software tool or tool suite capable of converting high level software languages like C++ or Matlab into a hardware description language (HDL). “In other words,” the explanation reads, “the topic is for a tool that helps describe what the computer hardware is, while the proposed technology is for software that controls how the computer hardware operates. These are two completely different types of technology.”
According to Zeidman’s complaint, HDLs are not specifically described by the language of the SBIR funding topic. “In fact, the topic description would not make sense with respect to HDLs because HDLs are synthesized, not compiled,” the complaint reads. “HDLs create a platform rather than get ported to a platform; and HDLs are used by hardware engineers, not software developers.” Bob Zeidman has designed and sold tools which work with HDLs, and the requirement would allegedly had been apparent to him if it was discussed in the topic description at all.
Zeidman next appealed to the U.S. Government Accountability Office (GAO) and he received a response from that office in mid-September of 2016 from an attorney at the U.S. Air Force Commercial Law and Litigation Directorate. This second explanation for the rejection from Air Force scientists on Zeidman’s proposal led to the conclusion that the proposal made no mention of the efficient utilization of multiple processing cores, which was “one of the main objectives of this solicitation.”
The suit filed by Zeidman notes that this updated reasoning was troubling because the project’s evaluators had changed the requirements from HDL to multicore processing, both of which are technologies that are not identified in the DoD’s request for proposals. Further, the proposal filed by Zeidman allegedly discussed how the company’s SynthOS tool is “processor agnostic,” enabling its use with existing processor architecture or C compiler architectures which could be developed in the future, thus meeting the criteria which was explicitly stated in the solicitation. “It appears that Defendant did not even read Zeidman Technologies’ proposal,” the complaint reads.
Zeidman received a third explanation for the rejection of its SBIR proposal after the issues with the second explanation were pointed out to the GAO’s attorney. This latest explanation focused on what evaluators felt was a fundamental difference between Zeidman’s approach, which “automates the process of creating the operating system so that the engineer can focus on writing the tasks that are specific to the devices,” and the solicitation’s objective, which “was to develop a tool that would automate the generation of optimized SW tasks/code for specific devices.” Further, project evaluators felt that the proposed innovation was unclear as SynthOS was already a patented product and there was no explanation to show how that commercial product would be improved upon for this project. According to Zeidman’s complaint, this was another instance of an explanation which relied on unstated criteria and ignored the evidence demonstrating Zeidman Technology’s qualifications.
During its communications with the GAO, Zeidman was informed that awards were granted in response to proposals filed by Colorado Engineering and EM Photonics for the same SBIR funding topic. In its protest of those bids, Zeidman notes that the websites of neither company reflect any expertise in either HDLs, and limited expertise in field-programmable gate arrays (FPGAs), both of which were supposed requirements for the proposal. Colorado Engineering in particular appears to have no software expertise as it is primarily a hardware design firm. While EM Photonics exhibits some expertise in software development, neither company exhibits any knowledge of accepting C-language code for a suite of applications and then outputing optimized code that can be compiled for a specific device.
In a telephone interview, Zeidman noted that he has applied for SBIR projects every once in a while over the past 30 years but he has never succeeded in getting one. By contrast, it appears that both Colorado Engineering and EM Photonics may enjoy some incumbency in being considered for SBIR grants according to information available online through the SBIR’s website. Colorado Engineering has received a total of 41 SBIR grants, either for Phase I or Phase II projects, earning nearly $14.5 million in grants for SBIR projects. Similarly, EM Photonics has received a total of 56 SBIR grants totaling just over $21 million. Both firms have earned the overwhelming majority of these government contracts for DoD programs.