TALENTSNAP

Hiring processes can be exhaustive, expensive, and inefficient. Over the years, there have been many iterations of software developed to improve the hiring process, but none have provided a truly revolutionary innovation in automating the hiring economy. The resources required to discover new candidates or job opportunities are costly, time-consuming and mentally exhausting. To further complicate things, web-based platforms that claim to be helping are actually centralizing our data and disseminating it without our consent.
Companies like Google, Facebook, Apple, and Microsoft have access to immense amounts of data, and construct multi-billion dollar business models around people's information. The more data you can feed AI models, the more purposeful they can become over time. The sheer amount of data these companies are processing, allows the aforementioned companies to be on the forefront of AI research. It is quite frightening that these companies know more about us than we know about ourselves. As it pertains to hiring, centralized AI systems can limit human potential by making decisions based on data that you never wanted disseminated. An open and centralized AI model can further perpetuate gender gaps and racial discrepancies leading to lower salaries and lesser opportunities.
Allowing a few centralized conglomerates to control our information can be potentially dangerous and can can lead humanity to a dystopic type of world, where Artificial Robots, and systems, run by a few conglomerates, control our thoughts, actions, and ultimate destinations in life.
Artificial Intelligence has the potential to automate every single industry in the world. But continuing to do this through centralized locations is dangerous because a handful of companies will be able to eventually control and manipulate the human race.
The economic model surrounding the
centralization of data ( of Machine Learning and Artificial Intelligence) is not aligned with the philosophical needs of the end-user, as people are no longer interested in having their data owned and controlled by large enterprises.
For this reason, data integrity needs to be ensured, and people rewarded for information that belongs to them by decentralizing and encrypting user data.
TalentSnap is going to revolutionize the hiring industry using Artificial Intelligence on the
decentralized, encrypted Blockchain to re-instill the power of data and monetization back into
the hands of its rightful owners, and ultimately automating the hiring process. TalentSnap is building a P2P Network and FitScore AI that will completely disrupt and automate hiring. TalentSnap aims to build powerful Zero-knowledge proof AI models on top of
decentralized and encrypted data; a new area of research that is immensely important.
TalentSnap will advance R&D efforts in Homomorphic Encryption, Federated Learning, and Proof-of-Identity, and combine it with the evolving FitScore technology to achieve true hiring automation.
Also, TalentSnap recognizes the failures of a singular centralized network being the primary source of professional data for candidate and employers around the world and is dead-set on exposing this flaw.
TalentSnap is a practical application of Blockchain to a specific set of business processes relevant to the staffing, recruitment, and job search industries.
BLOCKCHAIN TECHNOLOGY
Blockchain is the foundational technology of the network and provides a tamper-proof data structure, resulting in a shared public ledger accessible by everyone.
Difference Between Blockchain and Traditional Databases.
Traditional databases use client-server network architecture. Here, a user (known as a client) can modify data, which is stored on a centralized server. Control of the database remains with a designated authority, which authenticates a client's credentials before providing access to
the database. Since this authority is responsible for administration of the database, if the
security of the authority is compromised, the data can be altered, or even deleted while, Blockchain databases consist of several decentralized nodes. Each node participates in
administration: all nodes verify new additions to the Blockchain, and are capable of entering
new data into the database. For an addition to be made to the Blockchain, the majority of nodes
must reach consensus. This consensus mechanism guarantees the security of the network, making it difficult to tamper with.
Blockchain technology provides the following features:

1. Ethereum: Ethereum is a Blockchain-based distributed computing platform that runs smart
   contracts. It provides a decentralized infrastructure that is stable and secure. Ethereum features includes:
   a) Immutability – A third party cannot make any changes to data.
   b) Corruption and tamper proof – Applications are based on a network formed around the
   principle of consensus, making censorship impossible.
   c) Secure – With no central point of failure and secured using cryptography, applications are well protected against hacking attacks and fraudulent activities.
   d) Zero downtime – Applications never go down and can never be switched off.
2. Decentralized Control: Blockchains allow different parties that do not trust each other to share information without requiring a central administrator. Transactions are processed by a network of users acting as a consensus mechanism so that everyone is creating the same shared system of record simultaneously. Decentralized control eliminates many of the risks of centralized control. With a centralized database, anybody with sufficient access to that system can destroy or corrupt the data within. This makes users dependent on the administrators. Blockchain technology uses decentralized data storage to sidestep this issue, thereby building security into its very structure.
3. Immutable History: Blockchain databases are able to record data that is relevant now, but also all the information that has come before. Blockchain technology can create databases that have histories of themselves. They grow like ever-
   expanding archives of their own history while also providing a real-time portrait. The expense required to compromise or alter these databases has led people to call a Blockchain database immutable.
4. Performance: The way distributed networks are employed in Blockchain technology means they do not share and compound processing power, they each independently service the network, then compare the results of their work with the rest of the network until there is a consensus that something happened. Blockchain Technology is publicly verifiable, which is enabled by integrity and transparency.
   Integrity: Every user can be sure that the data they are retrieving is uncorrupted and
   unaltered since the moment it was recorded.
   Transparency: Every user can verify how the Blockchain has been appended over time.
5. Confidentiality: Blockchain is a write-uncontrolled, read-uncontrolled database. That means anyone can write a new block into the chain, and anyone can read a block in the chain. A permissioned Blockchain can be write-controlled and read-controlled. That means the network or the protocol can be set up so only permissioned participants can write into the database or read the database.
6. Smart Contracts: Smart contracts provide users with automated, low-fee transactions that provide a higher level of security than those that rely on standard localized databases. The smart contract provides a transparent procedure of showing the secured and decentralized AI transactions that will eventually automate hiring. Smart contracts will manage workflows between hiring parties and candidates with a
   level of accuracy not found in standard database technology.
   TALENTSNAP PLATFORM
   TalentSnap's platform is built on three layers namely:
7. Blockchain Layer
8. Artificial Intelligence Layer
9. Web Layer
   Blockchain Layer: TalentSnap is a practical application of Blockchain to a specific set of business processes relevant to the staffing, recruitment, andjob search industries. Blockchain allows for a shared, distributed, and encrypted ledger of transactions. TalentSnap is building a decentralized P2P candidate platform built upon the Ethereum Blockchain. The decentralization of data and the efficacy of Machine Learning will empower people to connect to jobs globally in a safer and more rapid manner than ever before. Blockchain technology will give ownership and control of data into the hands of the people; thereby allowing candidates to share information to their liking, and to be rewarded with TSC tokens through smart-contracts for participating within the platform. TalentSnap will build a platform that is decentralized, stable and secure, extendable, scalable and promote an exposed arrangement for user motivation.
   [8/14, 16:58] +234 806 808 2077 : Blockchain consists of the following features:

• Hashing Algorithm: Hashing is the process of mapping digital data of any arbitrary size to data of a fixed size. It is a process of taking some information that is readable and making something that makes no sense at all.
• Digital Signature: When signing a paper, a signature is appended to the text of a document. A digital signature is similar; personal data is needed to append to the document.
• InterPlanetary File System: IPFS (InterPlanetary File System) is a distributed file system technology based on DHT (Distributed Hash Table) and the BitTorrent protocol. It allows one to combine file systems on different devices into one, using content addressing.
• Asymmetric Encryption: Asymmetric cryptography is a branch of cryptography where a secret key can be divided into two parts, a public key and a private key. Public key can be given to anyone, while the private key must be kept secret.

1. Artificial Intelligence Layer: TalentSnap's Zero-knowledge proof Artificial Intelligence technology is being built to automate the hiring economy for both candidates and employers. The Artificial Intelligence technologies TalentSnap is building are scalable models on top of the Blockchain. Rather than looking at building AI models on readily accessible centralized data storages, TalentSnap's AI layer will consist of independent AI agents that work with data that is decentralized in independent nodes that are encrypted. In order to build independent AI agents that can make computations on decentralized and encrypted data, TalentSnap must follow three steps:
   a. The complexity and accuracy TalentSnap's current AI models need to improve, which will make more reliable recommendations on decrypted, centralized data sets.
   b. AI models, not controlled by a single administrator (i.e TalentSnap) must be trained to driving hiring decisions with data that is decentralized in many nodes.
   c. AI models, leveraging Homomorphic Encryption principles, will be able to compute hiring decisions on encrypted information.
   ARTIFICIAL INTELLIGENCE AND BLOCKCHAIN
   The "AI over blockchain" architecture for TalentSnap will be in part derived from
   OpenMined's architecture. The following step-to-step description of OpenMined's architecture will be used to describe TalentSnap's architecture,

• Initialization: The process starts by a company (e.g., TalentSnap) sending the description of an AI model to be trained as well as its public key to a smart contract. When the smart contract receives the model description, it builds a model according to the description and
  initializes the parameters of the model randomly. Since model description is public and the
  initialization is done randomly by the smart contract, there will be no room for manipulating the model in some way to store/steal people's information.
• Homomorphic Encryption: Homomorphic encryption is a technique used to make computations on encrypted data, without the need to decrypt it. This will allow participants, the candidates and employers on the network, to contribute data to the training of these AI models while still remaining encrypted to other parties. Once the smart contract builds and initializes the model, it sendthe model and the public key of the company to an oracle, which, stores the company's public key for future use, encrypts the model using its own public key, and sends the encrypted model back to the smart contract. Since the model has been encrypted by an oracle, only the oracle can decrypt it.
• Federated Learning: Once the encrypted model is sent back to the smart contract, it sends
  the encrypted model to people's data, each person runs the encrypted model on their data, the model improves slightly and gets back to the smart contract. Since the model is encrypted, people cannot steal the model. Since people's data is never leaving their device, their data is kept private. And since the model will not be decrypted until it has been trained on every person's data, one cannot reverse engineer the model and obtain sensitive information from it. For training of these AI models, TalentSnap will utilize core components of Federated
  Learning. Instead of people sending TalentSnap their data, TalentSnap sends them the AI model, the user runs this model on their secure data, and then the results proceed.
• Decryption: Once the encrypted model is trained on people's data through federated
  learning, the smart contract sends it to the oracle again, the oracle decrypts the model using
  its private key, encrypts the model using the company's public key and sends it back to the
  smart contract.
• Ending: The smart contract sends the trained model that has been encrypted with company's public key back to the company and the company decrypts it using its private key.
  OpenMined's architecture can be used for AI models that take in the data of one single
  person (and maybe some query) and make some prediction about that person. For instance,
  consider an AI-based medical assistant whose goal is to predict diseases based on medical and
  non-medical history of people. Once the AI model is trained, it can be encrypted and sent to a person's data to make a prediction about that person. Considering an application such as TalentSnap where for each search query of an
  employer, the AI has to be run on (potentially) all candidates to figure out which candidate is the best fit. In such a situation, a naive option would be to send the TalentSnap encrypted AI to every candidates, have them run the AI on their data based on the employer's search query,
  and return the best-fit candidates back to the employer. This option, however, will not scale
  when the number of search queries is high. It is also quite costly as every candidate should be
  paid for each search query (as they run the AI on their data) and quite computationally
  extensive.
  TalentSnap will build AI models based on two architectures, each consisting if two modules. They are:
  Architecture 1: The two modules are
  a) Data-free predictor: This module takes in a search query and a single person's data and predicts whether that person is a good fit or not. This module keeps no information about the users.
  b) Data-dependent blocker: A blocker can be viewed as a data structure which takes in a search query and quickly prunes a large number of candidates that are unlikely to fit the search criteria. To answer a search query, first the query is sent to the blocker and a small list of candidates that can potentially fit the search criteria is obtained. Then the model is only run on the candidates that pass the blocker.
  Architecture 2: The two modules includes
  a) Data-dependent predictor: This module creates an encrypted summary of each candidate which is stored on a decentralized database and can be used for making encrypted predictions about the user.
• Data-dependent blocker: A blocker stores data about candidates to be able to prune unqualified candidates for a particular search query.

1. Web Layer: TalentSnap will present its technology in a clean UI( User Interface). The user-facing component of TalentSnap platform will take the Blockchain and Artificial Layers and present them in purposeful application format to end-users.
   [8/14, 16:58] +234 806 808 2077 : TALENTSNAP UTILITY TOKEN
   A utility token provides users with future access
   to a product or service. TSC (TalentSnap Coin) is a utility token.
   TalentSnap is building a means of exchange that a large number of people globally can rally around. It is also, building a global hiring ecosystem that gives people and organizations
   access to connect with each other all around the world. The conversions and transfer of fiat currencies across borders is not scalable, nor efficient, and in utilizing a digital token like TSC, both the financial and frictional costs are dramatically reduced, if not eliminated completely. In the decentralized global economy, this is critical as a foundation for hiring. Using TSC token will allow users, both hiring parties and candidates alike, to set thresholds for automatic and manual exchange of funds, minimizing the frequency of transactions, and thereby the fees. Candidates and employers receive these tokens in exchange for their data, thus TalentSnap is rewarding them for their data in a way that presents real value, both for today and in the future.
   TSC is an ERC-20 utility token used to power the TalentSnap hiring ecosystem. TSC tokens will eventually have true utility for anyone within the labor force anywhere in the world. This token will allow one to be rewarded for his/her data, and in turn, will be able to use TalentSnap's FitScore AI technology which will automate the hiring or "getting hired"
   process. TSC tokens are a required unit of exchange within the TalentSnap network for
   employers and (potential) candidate members to utilize the FitScore technology which will
   eventually enable hiring automation.
   FITSCORE: FitScore is TalentSnap's rapidly evolving hiring AI that leverages highly sophisticated Machine Learning algorithms with the power to ingest hundreds of millions of unique indicators of candidates and employers. With terabytes of publicly available data to start with, and a roadmap of building decentralized labelled data, TalentSnap is able to generate a score from 1 to 100 on how well candidates and employers fit together. Multi-dimensional algorithms based on skills, culture, experience, in-depth personality variables, and dozens of other critical hiring metrics power this robust innovation.The technologies relevant to the FitScore includes the following:

• Natural language processing (NLP) techniques to extract meaningful information from job postings, candidate job descriptions, candidate summaries, and other social profile texts.
• Time-series analysis techniques to analyze the candidate's journey over the course of months and years and predict the next best company and role for them.
• Relational learning techniques to learn about the relationship between pairs of skills, pairs of titles, and also between pairs of skill-titles. Relational learning also helps combine data from several heterogeneous sources eg Resume, Homepage, LinkedIn, GitHub, Twitter, etc.
• Collaborative filtering and matrix factorization techniques to make better personalized recommendations to employers.
• Bayesian probability theory to make FitScore a self-adapting and self-improving technology which gets smarter and more precise as the users interact with it over time.
  TOKEN DISTRIBUTION
  TalentSnap will be creating a fixed supply of 403,333,333 tokens, of which 60% will be
  sold in the various phases of ICO. The remaining 40% will not be sold directly to investors, but will be distributed amongst advisors (10%) and marketing partners, including, airdrops and bounty campaigns (10%), as well as reserving a portion for company growth (13%)
  and the founding team (7%), ensuring the continued development of the platform.
  Public Sale: 46%
  Private Sale: 14%
  Company Reserve: 13%
  Advisory: 10%
  Marketing: 10%
  Founders Fund: 7%
  ALLOCATION OF FUNDS
  Personnel: 50%
  Marketing: 20%
  Tech: 12%
  Legal: 9%
  Partnerships: 5%
  Operations: 5%
  TOKEN DETAILS
  The Main Sale phase of the TalentSnap Token Sale will include the sale of up to 242,000,000 TSC and begin on October 1st, 2018 to November 30th, 2018.
  The maximum number of tokens allocated to ICO: 242,000,000 tokens
  First pre-ICO: 14,000,000 TSC
  Date of first pre-ICO: August 10th-August 31st 2018.
  Second pre- ICO: 242,000,000 TSC
  Date for second pre-ICO: September 1st-September 30th,2018.
  The TalentSnap ICO will be set with a soft cap of 2,000 ETH and a hard cap of 40,000 ETH.
  TALENTSNAP ROADMAP

1. Implementation of TSC Blockchain Layer: This layer will be built on Ethereum with highly customized smart contracts that will store the resulting state in a Blockchain. This will successfully facilitate transactions through the Blockchain, and transition towards true
   decentralization.
2. Decentralized P2P Network: TalentSnap is building a decentralized P2P system that aims to be highly efficient, secure and scalable. The P2P network will allow people to claim, or build their profiles, and facilitate connections between hiring parties and candidates. Within the P2P network, people will be able to claim their data from other social profiles, or build their own profiles from scratch.
3. TalentSnap's Artificial Intelligence Technology : TalentSnap will prioritize furthering Artificial Intelligence Research. Ultimately, TalentSnap's AI technology will become completely open-source. This will
   allow the Data Science and Software Developer communities globally to contribute at scale to
   solving such a complex problem.
4. Distributed Computing: In order to build a scalable decentralized AI and P2P network, Distributed Computing will be very essential. Eventually, these tasks will require heavy computation, which can be accomplished on a distributed computing network.
5. Applicationto Other Industries: TalentSnap believes that the technology being built will exponentially improve hiring, but can and will be applied to other industries as well. The technology can be applied to industries such as: Medical, Insurance, Logistics, Manufacturing, Legal, Finance and others.
   TALENTSNAP TIMELINE
   Q2 2017
   -TalentSnap is born.
   Q4 2017

• AI 1.0 MVP is released.
  Q1 2018
• TSC is conceived
  Q3 2018
• ICO begins.
  Q4 2018
  TSC is released.
  Q1 2019
• Processing Utility is implemented.
  Q3 2019
• AI 2.0 is released.
  Q1 2020
• Homomorphic Encryption.
  Q3 2020
• Federated Learning begins.
  Q4 2022
• TalentSnap becomes Open-Sourced
  Q4 2028
• AI reaches singularity.
  Q4 2035
• Autonomous Hiring