EXPERIENCE

EXPERIENCE IN DEPTH

CLC- Chen Lahav Consulting

I consult early-stage start-ups and ASX listed companies on technical implementation, innovation, and leadership in FinTech, adjacent and complementary domains. This entails validating the technical feasibility of business requirements under market and budget considerations, offering tactical and strategic technical solutions, working along c-suite executives and entrepreneurs, and contributing to the overall definition of the product roadmap while doing some coding, design, and code reviews. In addition, I consult pro-bono to technical and community initiatives set to promote the financial well-being of Australian woman and children and contribute to the software entrepreneur community by serving as mentor in Sydney Start-up Hub.

Hyper-Anna

Continuously generating insightful and actionable business insights is critical to any Financial agile organization but is currently limited to structured dialects, lengthy query development, lengthy gathering of business definitions, data talent in the organization and limited visualization. 

Hyper-Anna was set to ease this by enabling easier financial data exploration by developing an innovative SaaS product which enabled financial data exploration via a non structured, more natural language, advanced insight visualization, system generated insights and more. Hyper-Anna raised high funding, relative to the Australian market, from AirTree Ventures and others and was circa 70 people in its hight, circa half of which were R&D.

Using the product, one would upload a dataset and start asking questions. The results were presented in a way that enabled drill down into the insight, exploring trends and contributing factors. Once we learned that customers are either not sure what to ask or not asking the rights questions,  we developed a new capability to detect anomalies to be suggested as insights. The motivation behind this was to surface insights that are implicitly hidden to the user but deduced by the application following data enrichment and data introspection.

Leading the company operations operations helped me get acquainted quickly with the market in Sydney. Operations included leading the recruitment process, scaling the product, selecting and implementing an APM solution, setting a 3rd tier support and support procedures, managing the relationship with the different cloud providers and overall improving product performance. In addition, I took the initiative to implement security and quality processes. There was great emphasis on personal growth of employees via continues formal/in formal positive and constructive feedback. We prepared a matrix of expected skills and behaviours per each role and level and used it for performance reviews every couple of months employing additional known appraisal methods such as 180/360 degree feedback, ranking and others.

Barclays Investment Bank

Barclays was going through a major technical transformation when I joined the center of excellence in Tel Aviv. New regulations required that the bank puts aside collaterals for each investment in correlation to the calculated risk thus the importance of real time risk data was emphasized. In addition, the existing implementation, architecture and technology stack were outdated and did not scale. The existing technology was based on Microsoft stack (MsSql and SSIS) that struggled to process the vast amount of data. Data governance was lacking and data contribution SLA were repeatedly breached.

The bank had multiple risk data contributors (stock exchanges, financial institutions, etc.) push data in SOD (start of day) and EOD (end of day) and other times into different locations internally. File watching applications pulled the file system continuously and transferred the risk data, persisting it in MsSql for processing and transformation via SSIS. The SSIS transformed and normalized the data, producing the final data set and placing it in file system to be collected and visualized by risk data consumers. 

The new approach abstracted the implementation from the risk data providers and consumers by using language agnostic IPC such as messaging and web services. Risk data contributors were now instructed to contribute the data via rest api calls to web services that stored the data in column based Parquet files in HDFS. Data normalization and transformation was achieved via Spark Sql on top of Hadoop. Data consumers were instructed to request the data synchronously via rest or asynchronously via messaging. 

I noticed however that somethings were not attended to in the existing architecture and proposed implementation.

First, a language or protocol was required to enable data and business analysts to describe the business logic level dependencies between the numerous data sets. Second, a decision component was required to parse those definitions and verify that all required datasets were indeed contributed and it is possible to generate an accurate report. These were the capabilities my teams and I were leading.

MEDIA