A post-raise CTO with a hard integration and no senior hire in time
The Series A or B roadmap has an integration nobody on the team has shipped before, and the senior hire will not land in time. We take the integration as a defined outcome and hand it back documented.
02
A funded technology company with a performance, scale or reliability problem
The product works until it is busy. We measure the problem first, then fix the cause, and we say so when the answer is a smaller change than the team expected.
03
A regulated payments or asset issuer rebuilding for a new regime
MiCA, DORA, the UK regime, GENIUS, VARA, HKMA or MAS has changed what the system must record and prove, and the team needs engineers who have built for one of these regimes before.
04
A licensed fintech taking over a codebase from a previous vendor
The codebase arrived without the people who wrote it. We audit what is there, keep what works, and take ownership of the rest.
05
A founder with a co-founder gap
The company needs a CTO for one or two days a week to own the architecture, the roadmap and delivery with the team until a full-time hire makes sense.
06
A product agency without a senior backend bench
The agency has the design and the client, and needs a principal behind the backend for the length of the engagement without building that bench itself.
What we turn down
Hourly billing, a dedicated eight-person team as a product, and public-sector tenders scored on headcount.
What we build
Four things we build.
01
Backends and infrastructure
We build production backends and their cloud infrastructure. When an existing system struggles under load, we identify the bottleneck before proposing changes.
02
Payment rails
We build account ledgers, settlement services and card integrations with banks, issuers and processors. The scope includes transaction records and audit trails.
03
Identity and access
We integrate enterprise sign-on and build account and permission systems for people and software agents.
04
Distributed ledgers
We build systems that share and verify records across organisations. The Architecture Sprint establishes where a shared ledger is useful and where a conventional database is sufficient.
Work by Felag's founders and principals at Uddug and Gateway.fm, with prototypes and proofs of concept marked as such.
140k+
requests per second at peak in production, aggregate across clients, for Opera and 1inch
$1B+
locked in staking on the infrastructure we ran in 2025 and 2026, serving Lido, Gnosis, Stellar and Canton
9M+
users reported by Humanity Protocol, whose application layer we built
Each case gives its stage, what we built and the facts that add to it, with the notes and sources behind it.
Launched August 2024
01 / 10
Wirex Pay
We delivered the Wirex Pay platform end to end: the L2 chain, and the application layer, indexer, oracles and account abstraction behind its non-custodial debit cards.
Context
The chain was listed on L2BEAT, the public register of Ethereum layer-2 chains.
Sources and evidence
The engineering was built to support Wirex's MiCA alignment; we have not taken the system through an audit under MiCA. The L2BEAT listing and the partnership announcement below are the public record.
We designed the RPC proxy that served Opera and 1inch in production, with a cache layer and rate limiting.
Result
140k+ requests per second at peak, with 50 ms average latency on the heavy requests.
Measurement notes
Both figures are the peak observed in production, aggregate across all clients of the proxy, during 2021 to 2022. The latency is the average over the heavy requests only.
Gateway.fm, 2021 to 2022
Proof of concept, 2025
03 / 10
Identity for a global technology company
Two workstreams for one client. We designed and built account-abstraction wallets that people open with their company's single sign-on, and an identity system for AI agents on a major cloud platform.
Sources and evidence
The client is not named. The linked case study is published by Gateway.fm and covers the agent identity workstream only.
We built the backend API behind Infinex's Synthetix PerpsV3 integration. It takes in the PerpsV3 contract events and serves market data, order events and trading history to the app over REST and WebSocket.
Context
The scope was set by Infinex proposal XIP-9, Synthetix PerpsV3 API, approved in December 2023.
We built the application layer for Humanity Protocol, an identity network.
Result
A high-load application layer on the L2 chain, at about 1,000 transactions per second.
Context
9M+ users, as reported by Humanity Protocol.
Measurement notes
The transaction rate is the peak observed in production, aggregate across the chain, during 2024 to 2025.
Gateway.fm, 2024 to 2025
In use, 2024 to 2025
06 / 10
A shared prover for EY Nightfall, Polygon, Miden and Iden3
We built a shared prover used across several ZK protocols and L2 ecosystems, including EY Nightfall, Polygon zkEVM, Miden and Iden3.
Result
One prover for every protocol, with GPU and CPU capacity autoscaling on AWS spot instances.
Technical scope
The protocols served were Polygon zkEVM, Polygon Hermez, cdk-erigon, Miden and Iden3, including EY Nightfall. The cost effect of spot capacity was not measured, so no saving is claimed.
Gateway.fm, 2024 to 2025
Engineering leadership, 2024 to 2026
07 / 10
The Gateway.fm Platform department
The founders led Gateway.fm's Platform department and a second engineering department, accountable for the architecture and the delivery of their products.
Team
~25 people across the two departments.
Projects
Wirex Pay, Humanity Protocol, the shared prover, the oracles and Gateway Yield.
Project background
Uddug had grown to 15 people when it was acquired in June 2024, and the team joined Gateway.fm as its Platform department.
Gateway.fm, 2024 to 2026
Fractional CTO, ten months
08 / 10
DeepNode
Ten months as DeepNode's fractional CTO, designing its decentralised AI infrastructure through the January 2026 token launch and exchange listing.
Result
The network measured at about 5,000 requests per second.
Context
Listed in January 2026 at $140M+ fully diluted valuation at launch.
Measurement notes
The request rate is the peak observed in production, aggregate across the network, during the engagement.
Uddug built the whole NFT platform behind Arsnl.art and the NFT launches on it.
Result
Frank Stella's 'Geometries' released on the platform.
Project background
Uddug also handled the technical delivery for several other artist drops on the platform. 'Geometries' is the one named here, and sales figures are not stated.
Andrew builds and runs a crypto micropayment system and a network of faucets.
2015
Andrew Skurlatov and Mikhail Manko start working together and launch YourAdsMedia, a crypto ad network.
2017
The founders start ICEX, their own exchange, where the push oracle is first developed.
2018
The founders start Uddug on the ICEX core team.
2019
Andrew becomes technical lead and CTO of Snark.art, an NFT launch platform that later becomes a Uddug client.
2021
Andrew joins Gateway.fm as one of its first engineers.
June 2024
Uddug is acquired, and the team becomes Gateway.fm's Platform department.
2026
The departments close and the team leaves.
July 2026
Felag starts as the restart of Uddug, with Pavel Dodonov and Andrey Solovov as principal engineers and partners.
The full history: 4 eras and 25 projects, from a 2014 faucet network to Felag
Before Uddug
2014 to 2018
Andrew and Mikhail met in Limassol in 2014 and have built companies together since 2015.
2014 to 2015
Crypto micropayments and a faucet network
Andrew designed, built and ran a micropayment system and a network of faucets for Bitcoin, Litecoin and Dogecoin, earning about $10k a month at its peak.
2015 to 2017
YourAdsMedia
The founders' crypto ad network, product and platform built by Andrew and Mikhail.
2017 to 2018
ICEX
The founders' own exchange, with a team of 25. It was roughly 70 percent built when the investor withdrew in the 2018 crypto winter. The push oracle first developed for it later shipped for Wirex Pay, Infinex, LUKSO and Lumia.
Andrew Skurlatov and Mikhail Manko founded Uddug in 2018 on the ICEX core team, and it grew to 15 people. NFT launches by the team collectively raised $100M+.
2019 to 2020
Snark.art
Andrew was technical lead and CTO of Snark.art, an NFT launch platform for established artists, with a team of five. Snark.art later became a Uddug client.
Uddug built a mobile wallet for Pandoshi, as a consulting partner of Gateway.fm.
Gateway.fm
2021 to 2026
Andrew joined Gateway.fm in 2021 as one of its first engineers. In June 2024 Uddug was acquired and the team became Gateway.fm's Platform department; the founders ran ~25 people across two departments. In 2026 the departments closed and we left.
2021 to 2022
The RPC proxy for Opera and 1inch
A cache layer and rate limiting at 140k+ requests per second at peak, with 50 ms average latency on the heavy requests.
Andrew led backend engineering as Head of Backend, ~15 people, on a backend serving 100K+ requests per second at peak.
2023 to 2025
Presto and its fintech app kits
We led the Presto tokenised-finance product to revenue and built its stablecoin, RWA, DEX and payments blueprints.
2023 to 2025
The indexing platform
We built an EVM event-indexing code-generation platform that became the standard across most of the company's projects.
2023 to 2025
Oracles
A push oracle with multi-asset prices every 500 ms for Wirex Pay, Infinex, LUKSO and Lumia; a pull feed at about 1,000 requests per second and verifiable randomness for ApeChain. The figures are the peak observed in production, aggregate across clients, during 2023 to 2025.
2024
Wirex Pay
The L2 chain, launched in August 2024, and the application layer, indexer, oracles and account abstraction behind its non-custodial debit cards.
We built a shared prover used across several ZK protocols and L2 ecosystems, including EY Nightfall, Polygon zkEVM, Miden and Iden3, with GPU and CPU autoscaling on AWS spot capacity.
Ten months designing DeepNode's decentralised AI infrastructure, measured at about 5,000 requests per second, through the January 2026 token launch and exchange listing at $140M+ fully diluted valuation at launch.
Andrew presented a proposal for EVM compatibility on the Canton Network to its super-validators.
2025 to 2026
Gateway Yield
We ran Gateway Yield, the infrastructure behind $1B+ locked in staking, serving Lido, Gnosis, Stellar and Canton.
Felag
From July 2026
Felag was founded in July 2026 as the restart of Uddug. Pavel Dodonov and Andrey Solovov joined as principal engineers and partners.
What it costs
We sell outcomes, not hours.
Fixed-price sprints, builds and technical reviews. Monthly retainers for pods and fractional CTO work, with deliverables agreed in the statement of work (SOW). Prices in US dollars, excluding VAT.
01 / 05
Free technical workshop
A 60-minute session with a founder to understand your system, priorities and constraints. If an Architecture Sprint is useful, we agree its scope. Detailed specification and costed build options are part of the paid sprint.
Free
What you get
Your system, priorities and constraints, understood
A sprint scope, if a sprint is useful
Terms
Free
60 minutes
A reply within two working days of your email
Who delivers
A founder
02 / 05
Architecture Sprint
One to two weeks with a principal that end in a build decision: spec, architecture decisions, risk register and costed build options, with the fixed-price build or the pod that follows quoted from them.
From $10k
What you get
Spec and architecture decisions
Risk register
Costed build options
A build decision
Terms
One to two weeks
Paid in advance
Credited in full against a pod or fixed-price build signed within 30 days of the sprint's delivery
Who delivers
A principal
03 / 05
Fractional CTO
A principal in the CTO seat for agreed days each week, spread across Monday to Friday, until a full-time hire makes sense. The CTO owns the architecture and the engineering roadmap, runs engineering process and delivery with your team, and represents the technology to the board and investors.
From $4k a monthone day a week; two days and ad hoc days priced per engagement
What you get
Agreed days each week, spread across Monday to Friday
The architecture and the engineering roadmap
Engineering process and delivery with your team
The technology represented to the board and investors
Terms
Monthly in advance
One-month initial term
Who delivers
Andrew, Pavel or Andrey, never a contractor
04 / 05
Pods
The build after the sprint, on a monthly retainer with deliverables agreed each month in the SOW, and a team sized to the scope.
$10k to $60k a monthby team
What you get
From $10k: one senior engineer who plans the work with your team, with a principal reviewing the plan and the architecture weekly
From $25k: a principal part time with one or two senior engineers
Up to $60k: a principal full time leading two to three engineers
Terms
Monthly in advance
Deliverables agreed each month in the SOW
A missed deliverable triggers a scope and plan review, recorded in writing
Cancel with two weeks' notice
Fees for the unused part of a prepaid month refunded pro rata after the notice period
Who delivers
A principal, with senior and specialist engineers as the scope needs
05 / 05
Technical due diligence
A review of the system and the team behind it against a scope fixed in advance, for investors and acquirers who need findings they can act on.
Quote per scopefixed price, half paid on start
What you get
A written report on architecture, code quality, security, infrastructure and the team
Findings ranked by risk
Delivered one to two weeks from access
Terms
Fixed price
Half paid on start
Who delivers
A principal
Every build starts with a paid Architecture Sprint, and fixed-price builds are quoted from a completed sprint. Your engagement is led by one of our principals, identified in the statement of work. Any replacement requires your consent.
Free, with a founder. If a sprint is useful, we agree its scope.
02Architecture Sprint
A spec and costed build options.
One to two weeks with a principal, paid in advance and credited against the build.
03Decision
A build decision on a fixed scope.
Taken together on the sprint output: a fixed-price build or a pod.
04Build
The build, led by a principal.
A pod with deliverables agreed each month, or a fixed-price build.
Fractional CTO work and due diligence start from the same workshop and go straight to a scoped proposal.
Who we are
The people who lead the work.
Andrew Skurlatov
Co-founder and CTO
Based in the Netherlands
Andrew sets Felag's technical direction, leads the architecture on its engagements and steps in as fractional CTO. He has spent 15 years in software, ten of them in blockchain. He co-founded Uddug in 2018 and, after it was acquired, ran ~25 people across two departments as VP of Platform and Yield.
Mikhail looks after partnerships and deals at Felag, from the first call to the signed contract. He has spent 13 years running teams, products and operations. He co-founded Uddug with Andrew in 2018, saw it through its acquisition, and then headed the platform team for Gateway.fm's tokenised-finance product.
Pavel leads engineering on Felag's engagements. He came to software after a decade of scientific R&D and 20+ published papers. In nearly four years at Gateway.fm he went from Go backend engineer to Backend Lead, and led the architecture of the shared prover that several ZK networks ran on.
Andrey leads delivery on Felag's engagements. He started at Uddug, building smart contracts and payments architecture. At Gateway.fm he spent more than two years as Blockchain Dev Lead, shipping production systems for DeFi, oracle, L2 and identity clients, from the contracts down to the services behind them.
We bring in specialist contractors when a project needs them, always led by one of us.
Questions we are asked
Questions about working with Felag.
Do you bill by the hour?
No. Sprints and technical due diligence are fixed price, pods and fractional CTO work are monthly retainers, and the workshop is free.
Who actually works on our system?
One of the people above leads every engagement and is identified in the SOW. Contractors join for QA, DevOps and additional development when the scope needs them, always led by one of us.
How do we start?
With a free technical workshop with a founder. For a build, the next step is a paid Architecture Sprint. Fractional CTO work and technical due diligence go straight to a scoped proposal.
What do we keep if we stop after the sprint?
The spec, the architecture decisions, the risk register and the costed build options are yours. The sprint is paid in advance and credited in full against a pod or fixed-price build signed within 30 days of its delivery.
Where are you based and who do we contract with?
The contract is with FELAG LLC, a Georgian limited liability company with identification number 426561829 and registered office at Georgia, Gardabani Municipality, Poladaant Kari Village, 4th Street, Lane II, N 11. The founders work from the Netherlands and Cyprus.
How are deliverables accepted and scope changes agreed?
For monthly delivery pods, deliverables are reviewed and accepted with you at the monthly review, against the SOW, and scope changes agreed there go into the next month's SOW. For an Architecture Sprint or technical due diligence, the principal takes you through the output at delivery, and you accept it or return a written list of gaps within five business days. A fixed-price build is accepted milestone by milestone, against criteria agreed in the sprint, on the same five-day review after each demo. A review with no list counts as acceptance. Any change to a fixed-price scope is agreed in writing before work on it starts, with a fixed quote for its effect on the price and the timeline.
Do you work with regulated companies?
Yes. We built the Wirex Pay platform and the application layer behind its debit card programme, engineering that supported the client's MiCA alignment. We have not taken a system through an audit under MiCA, and we say so before a sprint starts.
Do you use AI in delivery?
Yes. We use AI tools and our own development framework. A principal reviews the work and remains responsible for the agreed deliverables.
Next step
Start with the workshop.
Tell us about the system and your deadline. We start with a free technical workshop with a founder, and reply within two working days.