Samed Duzcay

Samed Düzçay (also written Samed Duzcay) is the founder of Tori Labs, the company behind the on-chain yield protocol Tori Finance. He was previously the co-founder and chief executive of Infoset, an Istanbul-based customer-communications software company that was acquired by the cryptocurrency exchange BtcTurk in February 2023.^[1]​^[2]​

Early Life and Education

Düzçay studied at Boğaziçi University in Istanbul, one of Turkey's oldest institutions of higher education, founded in 1863. He completed a Bachelor of Science in Computer Engineering between 2015 and 2019, graduating with a grade point average of 3.80 and a reported class rank of second out of more than 100 students. He continued at the same university for a Master of Science in Computer Engineering, which he pursued from 2019 to 2021.^[1][3]

During his undergraduate studies he spent the spring 2018 semester as an exchange student in the Bachelor of Science in Computer Science program at The University of Texas at Austin, where he attended the Longhorn Energy Summit and a Science Sprint.^[3][4] At Boğaziçi he was a member of COMPEC, the university's IT club, and of BÜDAK, its mountaineering club, and was a founding member of BÜGAK, its futurism club.^[3]

Early Software Engineering Career

Düzçay's professional engineering work began at Pisano, an experience-management platform company founded in 2016 and headquartered in London. He joined as a Software Engineer in June 2016, working through June 2017 on full-stack web development using Ruby on Rails, AngularJS, CoffeeScript, HTML5 and LESS, in team-based development organized with Kanban and Scrum, and built, tested and launched the company's key web product, the Pisano Web Widget.^[3]​^[1] He was later promoted to Lead Software Engineer, a role he held into late 2018, during which he designed the architecture of, built in Scala and deployed Pisano Analytics, organized the company's first internal hackathon, mentored the software team, led the web product team for five months, and rebuilt the Web Widget's architecture using ES6/ES7, React, Redux and Redux-Sag, which served as the basis for later work with blockchain..^[3]​

In the summer of 2017 he worked as a Software Engineering Intern at Google in the Zürich area of Switzerland, where he built and tested the backend of a mid-size product feature and constructed a data pipeline to collect user statistics for analysis.^[1]​^[3] He has said that this period first introduced him to Ethereum and decentralized finance.^[2]​

Alongside his engineering roles, Düzçay served as Development Team Leader at the Boğaziçi University Entrepreneurship Center from September 2016 to June 2017, where he directed a team of six and built and deployed the center's web application, "Bright," using PHP (Laravel) and JavaScript.^[3]​^[1] Between September 2018 and February 2019 he co-founded Lisa AI, a small technology company for which he built parts of the mobile application in React Native and parts of the backend and big-data infrastructure using Hadoop, HBase, Phoenix and Spark.^[3]​^[1]​

Infoset

Düzçay co-founded Infoset in Istanbul in January 2020 together with Sercan Düzçay, serving as co-founder and chief executive officer.^[5][1] Infoset is a software company that provides cloud-based customer communication and sales tools, described as an all-in-one platform for managing support and sales channels from a single dashboard. Its product combines a cloud call center, live chat and no-code chatbot software, email ticketing, WhatsApp Business messaging, and drag-and-drop sales pipelines, with the stated aim of solving problems of integration, reporting and managing multiple products.^[1][5]

Less than six months after launch, Infoset raised a seed round from IT Task Force, a French information-technology company. The founders did not disclose the size of the investment but said it valued the company at $2.2 million.^[6] At that time Düzçay said Infoset had eleven paying customers with more than 100 users at those businesses, and planned to use the capital to grow its team and accelerate growth. The company's subscription pricing ran from $15 per month to $199 per month for its most powerful plan, which bundled CRM, call center and chatbot features, alongside a free plan aimed at startups.^[6] Its total reported funding was $110,000.^[1]

In February 2023, Infoset was acquired by BtcTurk, described as the largest cryptocurrency exchange in Turkey; the acquisition was publicly announced in early April 2023.^[1] Following the deal, BtcTurk's own customer-support services were to be carried out by the Infoset team, which offered communication-channel automation across cloud call center, email, chat and social media through what the company called "omni automation."^[5] Düzçay remained chief executive until September 2025 and then served as a board member of Infoset until March 2026.^[1]

Tori Finance

Düzçay founded Tori Labs in 2026 and is the founder of its product, Tori Finance, a yield protocol that brings institutional-grade, delta-neutral trading strategies on-chain.^[7]​^[1] The protocol went live on Ethereum in July 2026 and grew past $63 million in total value locked as of a subsequent product update; before launch it filled a $50 million pre-deposit cap.^[7]​^[8] Its seed round was led by Delphi Ventures, with RockawayX serving as the vault's risk curator and anchor liquidity provider, independently monitoring off-chain positions and capital flows.^[7]​

The strategy underlying the protocol is the currency carry trade: borrowing in a low-interest-rate currency such as the US dollar and investing in higher-yielding emerging-market currencies, while using foreign-exchange hedging to remain delta neutral so that currency movements are offset. Düzçay has described this as an approach historically executed mainly by large institutions such as pension funds and banks, which required tens of millions of dollars in capital and complex arrangements including local bank accounts, custody, tax registration and compliance approvals.^[8] Through on-chain tokenization, Tori is designed to open these strategies to ordinary users. Its architecture deliberately separates strategy design, which Tori controls, from trade execution, which is carried out by regulated institutional execution desks in each market; Tori defines mandates and risk parameters and pays those desks execution fees rather than trading itself.^[2]​

Users interact with the protocol through a two-token model. A user deposits the stablecoins USDC or USDT and mints a receipt token, then stakes that receipt token to earn yield; the resulting stake token is liquid and composable, and can be used as on-chain collateral to borrow further stablecoins and redeposit into the protocol in a "looping" strategy. The protocol's products are branded trUSD and strUSD.^[2]​^[7] Düzçay has stated that the protocol itself takes no leverage and is completely unlevered, with looping and leverage being user actions rather than protocol actions.^[2]​

On security, Düzçay has said the minting contract enforces one-to-one on-chain collateralization so that it cannot mint an unbacked token even if keys are compromised, that contract upgrades pass through a timelock currently set to 24 hours with no fast path, and that the protocol uses least-privilege roles so that separate keys are limited to minting, pausing or upgrading. He has cited operational key-management and infrastructure failures — rather than contract-logic bugs — as the cause of several recent decentralized-finance exploits, and argues Tori's timelocks, scoped roles and monitoring are meant to address that class of risk. For losses in the underlying strategy, Düzçay has described a layered defense that first absorbs losses from a protocol reserve fund capitalized by protocol revenue and, only if losses exceed that fund, adjusts the NAV of stake tokens proportionally; the protocol is designed to keep 2–3% of total value locked liquid on smart contracts for instant redemptions and to maintain credit lines and liquidity backstops, governed by predefined incident and drawdown playbooks.^[2]​

Interviews

Institutional Yield Strategies #01

On May 5, 2026, Samed Düzçay, founder of Tori Finance, appeared on the Delphi Digital podcast with Yan Liberman and Can Gurel to discuss Tori's structure, yield strategies, risk management, and use of blockchain infrastructure.

According to Düzçay, Tori was developed in response to differences between the size of the stablecoin market and the yield available through commonly used on-chain strategies. He identified the global carry trade as one strategy used in traditional financial markets, involving borrowing in a lower-interest-rate currency and investing in a higher-interest-rate currency. Tori's approach applies this strategy to emerging-market currencies while using foreign-exchange forward contracts to hedge currency exposure. Düzçay attributed the persistence of the yield differential, despite the use of currency hedges, to factors such as regulatory restrictions, central-bank intervention, local market access requirements, and limited participation in foreign-exchange forward markets. Düzçay described Tori as a strategy manager responsible for defining investment mandates, risk parameters, and the assets to be tokenized, while trade execution is conducted by institutional desks operating in the relevant markets. He stated that these counterparties include regulated institutions with established operations in emerging markets and experience executing similar strategies. Under the described structure, the execution desks operate within parameters established by Tori rather than independently determining the strategy. Düzçay identified market access and transaction volume as factors affecting the pricing available to participants in these markets. The protocol uses a two-token structure. According to Düzçay, users deposit USDC or USDT and receive a receipt token representing their deposit. Staking this token provides access to the yield generated by the underlying strategy. The staked token can also be used as collateral in lending markets, allowing users to borrow stablecoins and deposit them into Tori again. This process can be repeated as a leveraged position. Düzçay stated that Tori itself does not use leverage at the protocol level and that any additional leverage resulting from lending and redepositing is undertaken by individual users.

Tori's oracle design uses net asset value (NAV) for token valuation rather than secondary-market prices. Düzçay explained that this mechanism is intended to prevent temporary price movements on decentralized exchanges from directly determining collateral values in lending markets. Under this structure, the value used for collateralization reflects the reported NAV of the underlying assets. He also noted that the staked token has a defined unstaking period, which affects the use of leveraged positions and their unwinding.

The interview also addressed the reporting of assets and liabilities held outside the blockchain. Düzçay said Tori planned to use Accountable, a third-party attestation provider, to collect balance information from on-chain and off-chain sources through infrastructure operating in a trusted execution environment. Zero-knowledge proofs are used to allow aggregate information to be verified without disclosing the underlying position data. According to Düzçay, the resulting information includes reserves, liabilities, collateral, and other balance-sheet data. He identified strategy execution, including adherence to mandates and counterparty management, as an area that continues to depend on the conduct of external institutions.

Düzçay described several security measures implemented at the contract and infrastructure levels. These include one-to-one collateral requirements for token minting, a 24-hour timelock for contract upgrades, separate permissions for different administrative roles, multi-party production infrastructure, and two-factor authentication. He also stated that Tori's contracts had been audited by Sherlock and Nethermind and that a bug bounty program was planned. According to Düzçay, these measures were designed to address risks involving key management, administrative permissions, and infrastructure in addition to risks identified through smart-contract audits.

Düzçay outlined two categories of stress scenarios discussed during the interview: a secondary-market depeg without a loss in the underlying assets and an actual loss affecting the underlying strategy. In the first case, he described a situation in which market participants sell the token below its NAV while the underlying reserves remain unchanged. Under the proposed design, lending-market valuations would continue to reference NAV rather than the secondary-market price. He also described liquidity arrangements intended to support redemptions and market activity during such events.

For an underlying loss, Düzçay described a reserve fund funded by protocol revenue as the first layer for absorbing losses. If a loss exceeds the available reserve, the loss would be reflected in the NAV of the staked token. He also described procedures for different operational scenarios, including depegs, drawdowns, and security incidents, with predefined actions for each situation.

At the conclusion of the interview, Düzçay stated that Tori was preparing for an Ethereum mainnet launch. He described plans involving audited contracts, an Accountable dashboard providing proof-of-reserves information, institutional execution arrangements, credit facilities, lending-market integrations, and a two-token architecture. ^{[2] [10]}

Hackathons and Awards

Düzçay was an active hackathon competitor during his university years, placing in eight documented competitions between 2015 and 2017. His first-place finishes included the COMPEC Garaj Hackathon in November 2015 for a game about online information security; the Büyük Sıçrama IoT Hackathon in May 2016, among more than 20 teams, for WowDoor, a smart door lock, where he built the backend API on a serverless Microsoft Azure architecture; the Comodo Hackathon in December 2016, among 16 teams, for Wower, a cyberattack-protection app, where he built the desktop application in Electron, Node.js and AngularJS along with its Ruby on Rails RESTful API; and the TechSummit IBM Hackathon in February 2017, among 11 teams, for an anti-harassment smart bracelet, where he built the backend in Node.js, the machine-learning models in TensorFlow, and the full web application.^[3]​^[1]​

His other placements included second place at the TechSummit Hackathon in April 2016 for a Naive Bayes mail classifier, second place at the UNFPA Innovation Day in June 2016 for a micro-influencer project called "Goodwill Ambassadors," third place at the Invent & COMPEC Data Analytics Hackathon in May 2017 for building three TensorFlow machine-learning models to forecast sales, and third place at the Ford Otosan OpenXC Hackathon in November 2017, among 26 teams, for a real-time in-car data analysis application.^[3]​