Software Architect / Team Leader with experience delivering projects in machine learning, cloud and embedded systems.
Aktualisiert am 03.02.2021
Profil
Freiberufler / Selbstständiger
Remote-Arbeit
Verfügbar ab: 04.01.2021
Verfügbar zu: 100%
davon vor Ort: 100%
machine learning, c++, software architecture
Golang, python, tensorflow, keras, kubernetes, docker, ci/cd, cloud, tcp/ip, embedded
English
German
B1
Portuguese

Einsatzorte

Einsatzorte

Zürich (+50km)
möglich

Projekte

Projekte

1 Jahr 5 Monate
2019-08 - 2020-12

Development of a document understanding systems

CTO / Lead software engineer machine learning cloud architecture
CTO / Lead software engineer
  • Established engineering methodologies and practices for the team and helped drive product design.
  • Built the cloud application backend service architecture (microservices connected by a message bus) and the cloud deployment mechanisms (on Google Kubernetes Engine).
  • Built a python framework for AI training and inference using shallow classification mechanisms. This provides a common mechanism used by different intelligent services to train, evaluate and diagnose.
  • Built a Neural Network implementation in Tensorflow/Keras for table understanding; extracting rows and column structure from commercial documents such as invoices and delivery notes.
machine learning cloud architecture
own company [name on request]
Zürich
1 Jahr 10 Monate
2017-06 - 2019-03

Leading a team of engineers

Technical Leader
Technical Leader
  • Leading a team of engineers working on fact extraction for web documents, applying bayesian and neural network NLP techniques to text understanding. Responsible for the initial design of the project as well as day to day software development tasks.
Google
Zürich
1 Jahr 1 Monat
2016-04 - 2017-04

Development

Software Engineer
Software Engineer
  • Developed a tool to execute distributed data processing pipelines on Google Container Engine (GKE). Developed an unsupervised learning system to generate embedding vectors from sparse matrices. Applied generated embeddings to document retrieval and scoring of a web search engine.
  • Laserlike uses an internal search engine to retrieve web documents relevant to the user’s interests. Traditional keyword based retrieval systems perform less well when retrieving documents for abstract entities, for which relevant documents may not contain the exact keyword. The traditional approach to improve the results is to use synonym expansion of the query; I developed an alternative approach using unsupervised learning for searching documents based on vector embeddings and implemented an additional document retrieval backend that provides results in a hyper-sphere of the query vector.
Laserlike
Mountain View
4 Jahre 1 Monat
2012-03 - 2016-03

various tasks

Founder / Distinguished Engineer
Founder / Distinguished Engineer
  • Created the concept for the company, recruited the initial team and participated in the initial fund raising process.
  • The company built a network virtualization solution for data centers that provides per-tenant virtual networks in a way that is interoperable with network virtualization technologies used by service providers. The solution is based on an IETF Internet Draft that I published before starting the company.
  • In addition to the technical design of the product, I implemented the BGP engine which is the component that distributes the routing information to and from all the compute nodes.
  • After the Juniper acquisition, I was behind the decision to open source the software. This has proven to be a significant competitive advantage in a landscape where Contrail is mostly used in conjunction with IaaS solutions such as OpenStack and Kubernetes which are also open source software.
  • After the product became Generally Available (GA), I spent significant percentage of my time supporting customer deployments, often in person.
  • When Kubernetes became available, I built the integration between Kubernetes/OpenShift and Contrail.
own company [name on request]
Sunnyvale
1 Jahr 2 Monate
2011-01 - 2012-02

support in borg-master

Software Engineer
Software Engineer
  • Software engineer working in the cluster compute scheduler used by Google’s internal infrastructure (borg-master).
  • At the scale of a google cluster, a workload (a job instance) can take minutes to be scheduled and for its data to be loaded onto a executing machine. I implemented support in borg-master to be able to re-schedule and load the data required for this instance to execute when dealing with planned failures such as workload migration when a machine is rebooted. This minimizes downtime for instances and reduces the need of a service to have extra instances in order to deal with this (common) type of service unavailability.
Google
Mountain View
1 Jahr 5 Monate
2009-08 - 2010-12

Distinguished Engineering

Distinguished Engineer
Distinguished Engineer
  • Developed a prototype of a distributed database designed for control systems in emerging areas such as IoT.
  • At Cisco, my mission was to look at future direction for the routing protocols team. Routing protocols are distributed algorithms that discover the topology of networks and compute the path for network traffic. Given the heterogeneous aspect of the internet, routing protocols must be standardized and are naturally slow to evolve.
  • I focused my efforts on an attempt to use the know-how in distributed message passing systems gained in internet routing, to the fields that Cisco was betting on at the time: IoT and smart grid applications.
Cisco Systems
San Jose

Aus- und Weiterbildung

Aus- und Weiterbildung

4 Jahre 11 Monate
1990-09 - 1995-07

Computer Science

B.S in Informatics, Universidade de Lisboa, Lisboa
B.S in Informatics
Universidade de Lisboa, Lisboa

Kompetenzen

Kompetenzen

Top-Skills

machine learning, c++, software architecture Golang, python, tensorflow, keras, kubernetes, docker, ci/cd, cloud, tcp/ip, embedded

Schwerpunkte

machine learning

Produkte / Standards / Erfahrungen / Methoden

CAREER HIGHLIGHTS

  • Co-founded a company; lead engineering and technology for a cloud based AI solution to semi-structured document understanding.
  • Started a new effort within Google’s Knowledge team to extract facts from web documents.
  • Developed a distributed unsupervised learning system to generate embedding vectors from sparse co-occurrence matrices.
  • Founded a company (acquired by Juniper Networks).
  • Led the software development for Juniper’s EX product line of network switches
  • Was the lead software engineer for core Internet routing (BGP) at Juniper Networks.
  • Developed the Linux IPv6 stack.

Design / Entwicklung / Konstruktion

cloud architecture

Einsatzorte

Einsatzorte

Zürich (+50km)
möglich

Projekte

Projekte

1 Jahr 5 Monate
2019-08 - 2020-12

Development of a document understanding systems

CTO / Lead software engineer machine learning cloud architecture
CTO / Lead software engineer
  • Established engineering methodologies and practices for the team and helped drive product design.
  • Built the cloud application backend service architecture (microservices connected by a message bus) and the cloud deployment mechanisms (on Google Kubernetes Engine).
  • Built a python framework for AI training and inference using shallow classification mechanisms. This provides a common mechanism used by different intelligent services to train, evaluate and diagnose.
  • Built a Neural Network implementation in Tensorflow/Keras for table understanding; extracting rows and column structure from commercial documents such as invoices and delivery notes.
machine learning cloud architecture
own company [name on request]
Zürich
1 Jahr 10 Monate
2017-06 - 2019-03

Leading a team of engineers

Technical Leader
Technical Leader
  • Leading a team of engineers working on fact extraction for web documents, applying bayesian and neural network NLP techniques to text understanding. Responsible for the initial design of the project as well as day to day software development tasks.
Google
Zürich
1 Jahr 1 Monat
2016-04 - 2017-04

Development

Software Engineer
Software Engineer
  • Developed a tool to execute distributed data processing pipelines on Google Container Engine (GKE). Developed an unsupervised learning system to generate embedding vectors from sparse matrices. Applied generated embeddings to document retrieval and scoring of a web search engine.
  • Laserlike uses an internal search engine to retrieve web documents relevant to the user’s interests. Traditional keyword based retrieval systems perform less well when retrieving documents for abstract entities, for which relevant documents may not contain the exact keyword. The traditional approach to improve the results is to use synonym expansion of the query; I developed an alternative approach using unsupervised learning for searching documents based on vector embeddings and implemented an additional document retrieval backend that provides results in a hyper-sphere of the query vector.
Laserlike
Mountain View
4 Jahre 1 Monat
2012-03 - 2016-03

various tasks

Founder / Distinguished Engineer
Founder / Distinguished Engineer
  • Created the concept for the company, recruited the initial team and participated in the initial fund raising process.
  • The company built a network virtualization solution for data centers that provides per-tenant virtual networks in a way that is interoperable with network virtualization technologies used by service providers. The solution is based on an IETF Internet Draft that I published before starting the company.
  • In addition to the technical design of the product, I implemented the BGP engine which is the component that distributes the routing information to and from all the compute nodes.
  • After the Juniper acquisition, I was behind the decision to open source the software. This has proven to be a significant competitive advantage in a landscape where Contrail is mostly used in conjunction with IaaS solutions such as OpenStack and Kubernetes which are also open source software.
  • After the product became Generally Available (GA), I spent significant percentage of my time supporting customer deployments, often in person.
  • When Kubernetes became available, I built the integration between Kubernetes/OpenShift and Contrail.
own company [name on request]
Sunnyvale
1 Jahr 2 Monate
2011-01 - 2012-02

support in borg-master

Software Engineer
Software Engineer
  • Software engineer working in the cluster compute scheduler used by Google’s internal infrastructure (borg-master).
  • At the scale of a google cluster, a workload (a job instance) can take minutes to be scheduled and for its data to be loaded onto a executing machine. I implemented support in borg-master to be able to re-schedule and load the data required for this instance to execute when dealing with planned failures such as workload migration when a machine is rebooted. This minimizes downtime for instances and reduces the need of a service to have extra instances in order to deal with this (common) type of service unavailability.
Google
Mountain View
1 Jahr 5 Monate
2009-08 - 2010-12

Distinguished Engineering

Distinguished Engineer
Distinguished Engineer
  • Developed a prototype of a distributed database designed for control systems in emerging areas such as IoT.
  • At Cisco, my mission was to look at future direction for the routing protocols team. Routing protocols are distributed algorithms that discover the topology of networks and compute the path for network traffic. Given the heterogeneous aspect of the internet, routing protocols must be standardized and are naturally slow to evolve.
  • I focused my efforts on an attempt to use the know-how in distributed message passing systems gained in internet routing, to the fields that Cisco was betting on at the time: IoT and smart grid applications.
Cisco Systems
San Jose

Aus- und Weiterbildung

Aus- und Weiterbildung

4 Jahre 11 Monate
1990-09 - 1995-07

Computer Science

B.S in Informatics, Universidade de Lisboa, Lisboa
B.S in Informatics
Universidade de Lisboa, Lisboa

Kompetenzen

Kompetenzen

Top-Skills

machine learning, c++, software architecture Golang, python, tensorflow, keras, kubernetes, docker, ci/cd, cloud, tcp/ip, embedded

Schwerpunkte

machine learning

Produkte / Standards / Erfahrungen / Methoden

CAREER HIGHLIGHTS

  • Co-founded a company; lead engineering and technology for a cloud based AI solution to semi-structured document understanding.
  • Started a new effort within Google’s Knowledge team to extract facts from web documents.
  • Developed a distributed unsupervised learning system to generate embedding vectors from sparse co-occurrence matrices.
  • Founded a company (acquired by Juniper Networks).
  • Led the software development for Juniper’s EX product line of network switches
  • Was the lead software engineer for core Internet routing (BGP) at Juniper Networks.
  • Developed the Linux IPv6 stack.

Design / Entwicklung / Konstruktion

cloud architecture

Vertrauen Sie auf GULP

Im Bereich Freelancing
Im Bereich Arbeitnehmerüberlassung / Personalvermittlung

Fragen?

Rufen Sie uns an +49 89 500316-300 oder schreiben Sie uns:

Das GULP Freelancer-Portal

Direktester geht's nicht! Ganz einfach Freelancer finden und direkt Kontakt aufnehmen.