A roomful of game developers, a pair of armchairs, and a cocktail called Portal Punch. At a 2022 gathering during the Game Developers Conference in San Francisco, Junaili Lie sat down with Ian Proulx, co-founder of the studio behind Splitgate. The advertised subject was scaling from zero to millions of players. The drinks had the more entertaining name. The conversation addressed the problem that begins when a multiplayer game gets the audience its creators wanted.
Lie, who goes by Jun, had spent years working on that problem from inside game companies. By then he was doing it through AccelByte, the backend business he co-founded in 2016. More than 90 industry leaders attended the gathering. For an evening, some of the people responsible for keeping players connected were in the same room themselves.
There is a useful introduction to Lie in that scene. His work concerns the part of a game that people tend to notice when it stops working. A player wants to join friends, enter a match, and find yesterday’s progress intact. Someone has to make those apparently modest requests survive a launch, a new platform, and an unexpected crowd. Lie has made a career of taking them seriously.
From Indonesia to the machinery of multiplayer
Born and raised in Indonesia, Lie studied computing science at Simon Fraser University in Canada. His professional route led through Electronic Arts and LucasArts before Epic Games. These were places where online technology had to serve teams making and operating games, with the practical complications that come with both jobs.
An early colleague’s recollection offers a smaller-scale view of his working life. Vincent Lasmarias, writing a professional recommendation, remembered Lie returning to Guardian after a cooperative work term and taking responsibility for its database administration and architecture. Lasmarias credited his learning and work with helping an open-source system grow from tens of customers to thousands. It is a revealing assignment for a future founder: responsibility for the part everyone else depends on.
At EA, Lie held engineering and architecture roles; at LucasArts, he became an online games technical director. He subsequently served as director of online technology at Epic. There he helped assemble the team behind online infrastructure that would power services including the Epic Games Store, Unreal Engine Marketplace, Fortnite, and Epic Online Services. The distinction matters: his contribution was to the underlying team and infrastructure, rather than a claim to have made those games himself.
His game credits carry their own quiet record. FIFA Soccer 08 lists EA Sports World; NBA Live 10 lists EA Sports Online. Later entries include Infinity Blade III and Shadow Complex: Remastered. In the AccelByte years, the credits extend to Borderlands 3, Payday 3, Sunderfolk, and Wildgate. Read together, they trace a working life through sports, action, and multiplayer worlds. The recurring subject is the infrastructure underneath them.
The argument was about control
In October 2019, Lie published an essay called Owning Games Backend Platform. Its central concern was how studios could retain control of online technology without making every component themselves. He described two commercial approaches at the time: licensing the technology and a cloud arrangement he called lease-to-own. Both addressed the gap between needing a platform and having the people to build one.
He connected control with concrete choices: distributing across stores, carrying progression between platforms, adapting features, and retaining access to player data. A studio making several games could also use a shared backend across them. His argument treated infrastructure as part of a studio’s future, with consequences extending beyond the title currently on its desks.
The same essay dwelt on what happens after release. New features, player surges, deployment, and operating costs keep requiring attention. That is an engineer’s inconvenient addition to the launch party: the work continues after the cake. In Lie’s account, buying technology still needed to leave developers room to change how their own games worked.
Lie also described a wider ambition in 2021: a federation of independent studios sharing ideas, audiences, and resources while retaining creative autonomy. He floated server time as one resource they might exchange. It was an aspiration, rather than an announced organization. Still, it shows how he thought about the economics of independence. A studio could keep its own identity while gaining some advantages of working alongside others. That possibility gives his infrastructure project a social dimension: common technology might help individual teams cooperate, as well as help them put players in the same match.
A bigger cheque, the same problem
AccelByte announced a $10 million Series A in September 2021. Galaxy Interactive led the round, joined by NetEase, KRAFTON, and Dreamhaven. These were connections to investment and to game making: companies with their own reasons to care about how studios build and operate online services.
In May 2022, the company announced a $60 million Series B led by SoftBank Vision Fund 2, with Sony Interactive Entertainment participating alongside existing investors. Lie’s explanation emphasized what daily work with developer partners taught his team. The intended expansion included product and engineering capacity. For a business selling relief from technical work, the people doing that work remained central.
Financing raised by AccelByte. These figures describe company funding.
March 2023 brought a more direct expression of the accessibility argument: AccelByte Starter, a public-cloud offering introduced free in early access. It extended lessons from the company’s private-cloud business toward smaller developers. Lie’s stated concern was that rigid backend products could push infrastructure decisions into game design. He wanted teams of different sizes to have more room to make their own choices.
There is a practical tension in that ambition. Reusable services save work, while individual games demand differences. The founder’s challenge is to make enough of the machinery common without forcing the resulting games to resemble one another. The measure of the proposition sits in what a customer can actually ship.
Two launches and a founder’s relief
In July 2025, Wildgate and SUPERVIVE launched globally within days of one another. Wildgate came from Moonshot Games and publisher Dreamhaven; SUPERVIVE came from Theorycraft Games. Both used AccelByte, though their requirements differed. Wildgate needed cross-platform identity and custom matchmaking; SUPERVIVE required support for rapid session turnover and persistent progression.
Lie’s public reaction stayed close to the tools. He described investment in launch readiness, observability, profiling, and automated load testing. He singled out logging and traceability, which made issues easier to locate, and thanked the partner studios and his own team. A founder discussing an important week could choose many subjects. Lie chose the ability to see what was happening inside the system.
“It takes a village to launch a game...”
Junaili Lie · reflecting on the 2025 launches
The phrase suits the job. Matchmaking, accounts, progression, and operations meet in a player’s experience, however neatly an organization divides responsibilities. Lie’s comments expressed pride in preparation that reduced the need for urgent late-night troubleshooting. There is a particular pleasure in engineering work when the good news is how little drama it produced.
The work also includes helping studios change systems after a game is already live. In a public post about backend migrations, Lie described lessons gathered from moving several studios onto AccelByte. He pointed developers toward a guide intended to be useful with other providers and custom systems too. He also raised the possibility of sharing internal migration tools, potentially as open source, if people were interested. That was an offer to explore, rather than a completed release. It nevertheless put some of the attention on lessons a team could carry away and apply elsewhere. For developers responsible for a running game, the usefulness of advice depends on the details it helps them anticipate.
Still close enough to explain the tools
By August 2026, Lie was publicly identified as both CEO and technical director. AccelByte’s introduction of Teammate, an AI tool for developers using its services, invited people attending gamescom to meet him for a walkthrough. The founder was still being offered as someone who could explain the product’s present behavior and its direction.
Teammate’s proposed jobs were specific: inspect integrations, assess SDK upgrades, review infrastructure sizing, and retain project context. Its design required findings to carry supporting references; findings that failed validation were suppressed. The product also kept engineering decisions with the developer. These were early-access capabilities, with further checks and personas planned. They expanded AccelByte’s attempt to remove recurring integration work around a game.
That month also brought a partnership with Xsolla. Announced on August 19, it paired AccelByte’s backend services with payment and monetization tools, initially through a closed testing program. Lie linked the partnership to making technology available to studios of different sizes. The initial test group was intended to help shape the integration before a broader rollout.
Seen across a decade, the continuity is striking. Lie’s tools and commercial arrangements have changed, from the ownership arguments of 2019 to AI-assisted integration and combined payments. The practical question underneath persists: which repeated jobs can a studio hand over while keeping the freedom to make its game?
At that San Francisco gathering, scaling was something to discuss over a drink. Once players arrive, it becomes something a team has to do. Lie’s career has accumulated around that transition. The memorable part of a game may be a world, a contest, or an evening with friends. His chosen work is helping those evenings happen, then helping the next group log in.