The military writ large has struggled to retain specialized skills, especially tech talent, so the Marine Corps has turned to tech to identify key traits in recruits to inform management choices around retention. The new tech is aimed at enabling a broader transition in Marine Corps to better focus on retention — rather than just recruitment — and use analytics to inform decisions aroun Marines upon their entry into service.

“The intended outcome is to decrease [Marine Corps Recruit Depot] attrition and increase the success of applicants through at least their first term of enlistment,” Maj. Jordan Cochran, a spokesman for the corps’ Manpower and Reserve Affairs Department, said in an email.

The tech itself is being developed with John Hopkins University’s Applied Physics Lab and includes several algorithms that will work with data from the Marine Corps Recruiting Command, Cochran said.

The tools were recently discussed during a congressional hearing on talent management across the services. Lt. Gen. David Ottignon, the deputy commandant for manpower and reserve affairs, told lawmakers this month that the tools are in development and are part of a broader push for deeper data-based analytics.

“We … want to be able to round out our analytical systems,” Ottignon told lawmakers.

The tools are being developed with funding from the Department of Defense, Cochran said. Beyond just retention, the Marine Corps hopes machine learning can one day serve as a valuable tool to then go on to manage talent it does retain. But for now, the proof of concept remains in the “initial stages” of testing.

“Future uses are to expand the predictive capabilities to predict success through various milestones in a Marine’s career such as completing Military Occupational Specialty School, completing their first term of enlistment, reenlisting, promotion, and key billets,” Cochran said.

The post Marine Corps looks to machine learning for personnel retention appeared first on FedScoop.

Douglas Bush, who was sworn into the top Army acquisition job Feb. 11, said that increasing the speed and agility of how the Army buys software is a top priority. Achieving that will depend heavily on using recent authorities granted by Congress, including one that allows the Army to pilot a new way to purchase software outside of standard acquisitions practices.

“I believe we have the authorities we need — it’s a question of using them well,” Bush said Thursday during a call with reporters.

Bush said the main issue when buying software is the lack of flexibility in the way the Army is allowed to spend money. Traditionally, the Army is authorized to use specific types of funding — also known as a “color of money” — for certain types of programs, like research and development or production procurement.

The rigidity in that construction often slows down programs that cut across those areas, especially software-based tech, Bush added.

“I don’t believe the private sector distinguishes between [research and development] and the procurement of software, but we do,” he said. “Does that make sense anymore? I’m not so sure.”

Congress allowed the Pentagon and military services to test a new budget activity specifically for software called Budget Activity 8 in the fiscal 2021 National Defense Authorization Act. That’s a new lever Bush says he wants to pull, and he hinted that he hopes the flexibility it offers might increase.

“The funding might have to be more flexible,” he said.

Bush used to work in Congress, most recently as a senior staff member of the House Armed Services Committee. He stressed that he plans to include Congress in key decisions he makes and will consult with members closely on budgetary matters. His plan for boosting software acquisition agility has yet to be finalized, he said.

“I can’t say I’ve got a master plan, but I want to develop a plan … to get us better than we are,” he said.

The post Army needs to better use recent software authorities, new acquisition leader says appeared first on FedScoop.

In a letter to Shareholders dated February 14, Alex Karp noted that Silicon Valley companies selling data products to federal agencies face complex moral questions, and that there “are no goods or services or government today without software.”

“We understand that all technology, including ours, is dangerous, and that software can be used as a weapon,” Karp wrote.

The company has come under scrutiny for partnering with the government, especially under the Trump administration. Karp has routinely defended that work, saying that it is the duty of his company — and of all companies — to work with the U.S. government if they are to also deliver tech capabilities to adversaries.

Most recently, the company won a $116 million Army contract that extended a third year on its work with the Vantage data platform.

Despite the comment of the lethal applications of its tech, the company has been trying to branch out from the military and government space.

“Most of our customers are commercial, most of our revenue doesn’t come from intel and defense,” Karp said at the Reagan National Defense Conference in December.

The company’s latest financials show that in quarter three of 2021 a majority of revenue still came from government contracts. The government revenues have included health care work from agencies like the Department of Health and Human Services, despite higher growth in its commercial sector work, according to investor presentations.

The DOD has been focused on developing its own data platform, Advana. The platform was cited in memos from the deputy secretary of defense as the DOD’s senior leader go-to place for data curation.

“The [Advana] platform is the single enterprise authoritative data management and analytics platform for the Secretary of Defense, Deputy Secretary of Defense, and Principal Staff Assistants (PSAs), with inputs from all DoD Components,” Deputy Secretary of Defense Kathleen Hicks wrote in May 2021.

She added all other data platforms need the sign-off from the CIO or chief data officer.

The post Palantir CEO says lives have been ‘saved and taken’ with its software appeared first on FedScoop.

Oracle cloud is now authorized to host top secret/sensitive compartmentalized information (TS/SCI) and special access program data, according to the company. Its accreditation applies to Air Force for now.

The company built special “National Security Regions” that are walled off from the open internet to ward of adversaries attempting to access sensitive data hosted in the cloud. Oracle said its systems also include strong encryption and security controls and in-depth auditing.

“DoD assessors granted the Authority to Operate (ATO) which will enable secure processing of some of the Air Force’s most sensitive data,” the company said in its press release.

Oracle is already used across the DOD at other security levels known as “Impact Levels.” Some of its tech offerings already being used include data base cloud services and data analysis tools, the company said.

Obtaining an ATO is often the most time consuming part of deploying a new technical system or piece of software across the DOD. It is a formal declaration by a designated authority that explicitly accepts risk to agency operations.

Microsoft and Amazon Web Services have previously said they have also been accredited to host highly sensitive data.

The post Oracle cloud authorized to host top secret data for DOD appeared first on FedScoop.

The most recent cohort of participants started taking an “AI 101” course in early February through a partnership with the Massachusetts Institute of Technology while another recently entered an AI coding bootcamp. The range of educational offerings from the AI-accelerator is designed to eventually be transitioned to other DOD institutions for tens or even hundreds of thousands of people to learn about AI, Greg Allen, the JAIC’s head of policy and strategy, told FedScoop.

“We are running training pilots to really test,” Allen said. “We partner with the broader department of defense … to help them deliver education materiel at scale.”

Allen added that “ultimately there is going to be hundreds of thousands of folks” getting some form of AI training, out of the DOD’s nearly 3 million employees and service members.

The work stems from a congressional mandate for the JAIC to develop an AI workforce and education strategy in the fiscal 2020 National Defense Authorization Act. The JAIC is now implementing that strategy through educational pilots, Allen said.

In that strategy, the JAIC identified six archetypes of AI learner: Lead AI, Drive AI, Create AI, Employ AI, Facilitate AI and Embed AI. Each type of DOD employee needs a different level of detail on AI, so the JAIC is leaning on different platforms to teach them.

For general and flag officers at the highest ranks of the military, Lead AI is an in-person seminar on the basics of what AI can do and how it will impact the capabilities they oversee. On the other end of the spectrum is Create AI, a group of coders that will be developing machine learning models for the military and need specialized training in developing machine learning models.

One of the adjustments the JAIC has made to its offerings for the Create AI category is a new coding bootcamp on the Python coding language that is often used to develop AI.

“These are folks who actually need all the skills to meet their current and future operational needs,” Allen said.

By 2023 the JAIC hopes to have all these lesson plans transitioned to other organizations, like the Defense Acquisition University or the Air Force’s Digital University.

“The No. 1 thing that brings the most joy to us is when we hear back from past participants … they put what they learned to practice in their jobs,” Allen said.

The post JAIC piloting artificial intelligence education for DOD appeared first on FedScoop.

The firm now has a new managing director, Jake Chapman who spoke with FedScoop in his first interview. Chapman is still working to get funding from Congress to invest in startups, but if given $50-$100 million, he says the Army Venture Capital Corporation could be a critical part of getting more startups and emerging technology in the hands of troops.

“I think our organization can play in that liminal space where there is a working product, and there is a buyer but most [venture capital firms] won’t touch the company, but we could step in and say alright we will bridge you for the next 18 months to two years until that contract comes through,” Chapman told FedScoop.

The Army Venture Capital Corporation’s (AVCC) first iteration was a failure, investing in companies that deflated during the 2008 recession and not building the goodwill it needed in the valley to be a trusted partner. Or a least, that’s Chapman’s diagnosis and exactly what he wants to avoid.

“The first ten years were not super successful from a financial perspective,” Chapman said. “Army circa 2002 did not really know what was going on in venture capital.”

Chapman said he brings the type of experience the Army needs to succeed this time around. He is a trained lawyer and has led venture firms in the valley focusing on “deep tech” like aerospace and biotech.

“The number of people that looked hard at the defense tech space … is a small group of people. I am one of them,” he said.

The problem for many startups that want to work in the defense space is that money is relatively easy to find in small, short bursts but major contracts take years to land. Without big contracts in hand, venture firms are often reluctant to bet on a company that wants to work with the military. That dynamic has often pushed startups to focus on commercial clients that can bring them revenue.

Chapman’s solution: give startups a few million dollars as a signal to investors that the Army is really interested. The power of the AVCC will be in the “signal” they send to the investor community with their cash. Couple its investments with a strong network in the valley, Chapman says AVCC can be a critical part in bridging the so-called valley of death.

In addition to bridge funding, Chapman wants the AVCC to focus on deterring adversarial capital from funds with ties to countries like China and Russia and doing what he called “ecosystem mapping,” or building data sets on the startups and private funds that want to work with DOD.

“We want to push and pull technology into the department,” he said.

Chapman

Chapman has more than a decade of experience in the valley and working with venture investments, work he says gives him the perspective to know what venture firms need to hear from companies seeking a new funding round.

His pivot from the commercial world to DOD was largely driven by a desire to apply his skills to national security problem sets. His father and members of his family served in the military and he felt his skills could best be applied to financing the tech needed to win the next war.

“I think one of the best ways of doing that is investing in national security tech,” he said.

Lobbying Congress

Chapman continues to work on building support for the orginization. While the authorities for AVCC remain on the books, it doesn’t have money. He hopes to bring on six people to help run the shop but so far his top priority is lobbying Congress for money.

He said he has seen support from both sides of the aisle and chambers for the idea. Similar initiatives have been raised to appropriate $100 million for an “innovation” fund.

“This is a totally unique tool in the DOD tool set and it would be a shame to let it go away,” he said.

The post The Army is trying to jump-start its venture capital arm appeared first on FedScoop.

The Department of Defense has sponsored hackathons before, typically focused on security bugs. But, “Bravo 0” was the first secret-classification hackathon with 80 hackers working on a so-called “air gapped” system at Nellis Air Force Base. The event yielded several new applications that Stuart Wagner, the Air Force’s chief digital transformation officer, said that senior leaders across the Air Force are interested in using in real-life scenarios.

“It felt like we’re beginning to miss the information age,” Wagner said of the inspiration behind the event. He invoked the Air Force’s founder, saying “we needed a Billy Mitchell moment, a project B moment” to spark innovation across the department.

The long-term goal of the event was to show more senior leaders how developers can solve problems. With senior leaders bought into the power of software—especially software crafted at the hands of airmen—Wagner said the Air Force will have more political power to attack the real problem: the Authority to Operate, or ATO.

The ATO is the largest speed-zapping bump in the road for developers. It’s a required step to getting code on Air Force systems to ensure there are not gaping security holes, but critics of the model say it’s outdated and to focused on check lists compliance over than actual security.

“ATOs are heavily oriented toward the assessment of doing something,” he said. “The [risk management framework] never asks what the risk of delay is.”

A more immediate goal, Wagner said, was developing ways for airmen to test unauthorized software on real data to know what is worth sending through the ATO process vs. what prototype apps only works in a synthetic environment with made-up data.

“Now I can prioritize where I can spend my ATO fires,” he said.

Wagner was able to take open source software and use it on real data because he built an air-gapped environment with developers bringing in code on DVD discs. He also allowed developers to use their own virtual machines while working a platform initially developed by the Defense Advanced Research Projects Agency (DARPA) that the Air Force picked up called “STITCHES.” The Air Force also used the Joint Artificial Intelligence Center‘s Joint Common Foundation in their tech development stack.

Beyond the tech, new cultural practices the Air Forced used was a relaxed dress code, 24-hour operations for coders to choose their own hours and a short six-day window to complete projects.

“It’s the best ideas that wins, not the rightest rank that wins,” Wagner said.

No specific problem areas were assigned, coders were just given data and told to make solutions to the problems they saw, Wagner added.

The post Air Force hackathon puts real data on open source code appeared first on FedScoop.

Google Cloud says their offering adheres to zero trust principles and use advanced machine learning-based firewalls to ensure the security of cloud and application access by DIU. The production contract could lead to broader adoption of the tools and services by the Department of Defense. DIU previously ran pilots with Google Cloud, Zscaler and McAfee Public Sector for multi-cloud management services, with Google Cloud coming out on top.

“In today’s new cybersecurity paradigm, it’s critical that government agencies see the benefits  of adopting a zero trust security strategy and have the option of selecting more modern, cloud native solutions that meet their unique needs,” Lynn Martin, vice president of North America public sector at Google Cloud, said in a statement.

The secure cloud management service is a container-based offering for application access and monitoring built on Google Cloud’s “Secure Application Access Anywhere” offering and uses its Anthos platform. It’s a new “standards-based solution” that Google Cloud says will serve as a scalable alternative to DOD’s network boundary points. The idea is to enable management of hybrid cloud and multi-cloud applications via Kubernetes, an automated container orchestration tech.

The contract award reinserts Google Cloud into the DOD’s cloud ecosystem after Google had previously stood down from several major defense technology projects in recent years. It had backed out on ethical grounds of the the Pentagon’s Joint Enterprise Defense Infrastructure (JEDI) cloud contract, which has since been cancelled. Google also walked away from a major artificial intelligence contract, Project Maven, after employees protested the project.

The post DIU selects Google Cloud for cloud management contract appeared first on FedScoop.

How does the Navy plan to keep watch for everything from hostile Iranian drones to an aggressive Chinese posture to rogue pirates?

Robots.

At least, that’s the plan for the Navy’s Middle Eastern contingent, the Fifth Fleet, where an acute nexus of manning the watch over a hostile region is colliding with diminishing resources. At the heart of the effort — to deliver unmanned systems to extend the vision of Naval Forces Central Command — is a group of 21 sailors, many of whom are reservists, that are catching the eyes of senior leaders in the Pentagon.

Task Force 59 launched in September and already is testing tech in the water, collaborating with foreign militaries and earning shout-outs from the Navy’s top sailor. (Chief of Naval Operations Adm. Mike Gilday said in January the task force is “catching fire.”) It’s now in the midst of helping lead an exercise with 80 unmanned systems from 10 different countries.

The task force’s goal is to deliver “deterrence by detection,” meaning having unmanned systems chock full of sensors to increase the eyes and ears sailors have across the oceans. It’s a technical challenge that requires everything from building robots, artificial intelligence models and complex IT systems to orient unmanned systems along with those that are manned.

But, is a group of sailors — nearly half of whom are reservists on temporary assignment — an example of success the Navy can scale to meet its needs? Or is it just another experiment that will only yield one-off results?

Building a team

The task force’s commodore, Capt. Michael Brasseur, says it is less about the size and more about the experience of his team. By bring on reservists, Task Force 59’s roster includes the CEO of a 1000-person cybersecurity company; a D.C. think-tanker working toward a Ph.D. from Harvard and an advisor who was one of Snapchat’s earliest employees.

“This is not the B team,” Brasseur said in an interview.

“I was floored by the caliber of people,” Cmdr. Tom McAndrew, the task force’s No. 2, said in a separate interview.

But the model that has brought in such a high caliber of people might mean that it is harder to replicate and eventually scale the successes it reaches. Ten of the 21 members are reservists, including McAndrew, meaning eventually they will go back to their day jobs outside the Navy.

“We certainly recognize the importance of developing a more sustainable staffing model and have already begun working toward this transition,” NAVCENT spokesperson Cmdr. Tim Hawkins said.

What’s under the hood?

The task force so far has put a saildrone and other vessels with unmanned capabilities in the water. But what Brasseur and McAndrew are most excited about — but at this point have yet to share details of — is the tech stack and IT that power the unmanned systems.

“While I think the boats and physical things get a lot of recognition, one of the most difficult things to do is the IT environment,” said McAndrew.

Brasseur said the Task Force has a company on contract to build the user interface. He wouldn’t specify what firm but promised it has shown great PowerPoint slides so far. Both Brasseur and McAndrew were shy to share what they have built as the IT is still in a test and evaluation mode, saying so far the most important thing they have is the team of sailors building and using it.

“Never before have I seen so few do so much, so fast,” Brasseur said. “Everyone is focused on the robots, but we are building a data infrastructure to support a place where we can exploit with AI and ML.”

What does this mean for the rest of the Navy? 

“I like the idea they are bringing in smart new people. I am sold that would work. But I think the problem with that is it’s not a scalable model,” Bryan Clark, a senior fellow at Hudson Institute, said in an interview.

Defense expert Melanie Sisson, a fellow at the Brookings Institute, warned against experimenting for experimentation’s sake. Beyond finding the right technology, she stressed the importance of finding the right use cases for the right technology for the task force’s success.

“The military is struggling to understand functionally what it will need to do,” Sisson said.

And where Task Force 59 sees senior leader buy-in, Clark warns of a “shell game” for what success means. How much the success of Task Force 59 will mean in success for the Navy can’t be known — but both Clark and Sisson warn that the key will be translating tech to the needs of operators.

“The Navy has a history littered with these programs that have come and gone,” Clark said.

It’s a point Task Force 59 says it can’t agree with more. The goal of the experiments, Brasseur said, is to get as much feedback as possible from operators — getting “kit in hands,” as he said.

“We have real problems and we are focused on solving real problems,” Brasseur said. “We are not tinkering.”

The feedback the task force is hoping to generate is also being shared with a broader set of technologists across the Navy. The Navy’s Joint All Domain Command and Control (JADC2) contribution, the Project Overmatch office, is clued into the work being done in the Middle East along with other unmanned task forces. JADC2 is the military’s new strategy to combined data from sensors across the battlefield, be they unmanned systems or manned sensors, into a military Internet of Things.

So far, much of the feedback Brasseur said he has gotten has been positive.

“We are trying to build real capability — we don’t have time or resources to do anything but that,” he said.

The post Task Force 59: The future of the Navy’s unmanned systems or a one-off win? appeared first on FedScoop.

Katie Arrington has been on leave from the department since May, following allegations she disclosed classified information. In October she filed a lawsuit against the DOD and National Security Agency in response to the allegations relating to her suspension.

The lawsuit was settled in late January, and in her resignation letter Arrington wrote that the allegations were false and political motivated.

“The suspension of my clearance was, in my perception, a politically influenced action driven to silence me,” she wrote.

Tuesday morning, Arrington announced her bid for Congress.

Before joining the department in 2019 Arrington ran for Congress to represent South Carolina’s first congressional district in 2018, winning the Republican primary but losing the general election. Before, she was a Republican member of the South Carolina State Legislature.

Arrington came to DOD as a “Highly Qualified Expert,” later earning the same title of chief information security officer of acquisition and sustinament as a member of the senior executive service. The CISO position she held was eliminated prior to her Monday resignation.

Arrington won praise for her work promoting CMMC, the department’s program aimed at increasing the cybersecurity of defense contractors by verifying their compliance to a range of security controls. She used her political skills to stump for CMMC, making cross-country trips and appearing weekly on zoom calls to speak in support of the program.

Some key members of the CMMC ecosystem bristled with her style and found her experience on cybersecurity policy lacking, a tension that at times strained her relationship with the CMMC-Accreditation Body that she helped create. After her suspension the department announced a significant reduction in the scope of CMMC and the program she built.

In her resignation she stated she had intended to file a hostile work environment claim against Deputy Assistant Secretary of Defense for Industrial Policy Jesse Salazar prior to her suspension.

The DOD and NSA declined to comment on her resignation.

The post Katie Arrington resigns from DOD, announces bid for Congress appeared first on FedScoop.