Expert profile: Dr. Alexandre Girardi

---

CURRICULUM VITAE

ALEXANDRE GIRARDI

alexandre-girardi@neuf.fr
Phone: +33 3 2728 7088 or +32 4 8586 5517

Personal Data:

Name: Alexandre Girardi
Status: Married
Birthday: March 10th, 1965
Nationality: Italian and Brazilian

Introduction:

I have supervised small and medium international teams of developers and 
researchers. I have experience dealing with research and development, sales 
managers, human resources, procurement, production lines, involving or not 
multiple partner companies. I have a strong background in specific fields as 
speech recognition and on management coordinating/writing/submitting/defending 
viability studies, writing strategic planning, business planning, business 
development, following them up and keeping technology exchange programs with 
different companies, universities and research sites in different countries. 

The master and doctor courses in advanced computer science I did, enabled me to 
work with state-of-the-art pattern recognition and synthesis algorithms, memory 
management, content management and data structure for fast and efficient access 
to data and knowledge retrieval. Memory management was designed and optimized 
in many different ways with regards to storage space, search and recovery 
methods. With the resulting research was possible to query huge libraries of 
audio and video automatically and efficiently for high demand and real-time 
applications. I conducted the work as program manager using diverse languages 
as English, French, Spanish, Portuguese and Japanese to mention a few, not only 
for documentation, but also for communication among team members and 
elicitation of user requirements. With this multi-lingual and 
multi-disciplinary team, the importance of pear reviews, feedback, different 
optimal working conditions (isolated and group activities) was developed 
through repeated cycles of results and processes planning, supervision, 
measurement and analysis. Consistent, continuous and motivational training 
programs boosted these cycles.

In order to support research in speech recognition, besides direct revenues 
from clients, I have applied and successfully granted national research funds 
(over 3 million dollars), carried out in private company owned and directed 
companies and institutions. This work is easier, since I have being also on the 
other side to the table as an evaluator of European projects under the 6th 
Framework Program (FP6).

Additionally one of my strongest points is the subcontracting of third 
companies, which gives a good response to high changes in demand from clients. 
Careful requirements specification and an often and close client interfacing 
result in smooth and consistent revenue flow. Clients in this case being both 
third party companies, investors, internal teams and of course the final 
customer.

I have worked with public (government) as well as with private companies, with 
hiring processes and career development, binding programs, research training 
and development, teaching leadership spirit to the team, dealing with 
investors; I like to building up motivation and designing the technology that 
customers are demanding and expecting for the near future. 

Currently, at Multitel, I am coordinating several groups in different domains 
such as speech recognition, synthesis and dialog systems. My main interests 
areas are management, speech processing and dialog management.

Sincerely,


Alexandre Girardi


Educational Record:

Doctor NAIST (Nara Institute of Science and Technology - Japan)
 Area: Computer Science - Acoustic Modeling and Voice Morphing
 Period: Oct/1996 until Mar/2001

Research Student at NAIST (Nara Institute of Science and Technology - Japan)
 Area: Speech Recognition
 Period: Apr/1995 until Sep/1996

Master ITA (Technological Institute of Aeronautics - Brazil)
 Area: Context Independent Recognition using Neural Networks
 Period: Aug/1990 until Dec/1994

College Electronic Engineer, ITA (Technological Institute of Aeronautics - 
Brazil)
 Period: Mar/1984 until Dec/1988

Management Many institutions and FGV (Fundacao Getulio Vargas - Brazil)
 Period: 500 hours of ISO 9001, CMM and management among others. 1994 until 
2003.


Publications:

L. Borges, A. Girardi, "Speaker Verification System for Security Systems Using 
Eigenvoices Technique", I2TS?f2003 - Brazil (2003-11)

L. Borges, A. Girardi, "Speaker Verification System Using Eigenvoices 
Technique", 2nd International Symposium of NAIST-IS 21st century COE program, 
Japan - invited talk-(2003-10)

I. Sanches, A. Girardi, "Multi-dimensional Filtering for Speech Enhancement via 
Microphone Array", 2nd International Symposium of NAIST-IS 21st century COE 
program, Japan - invited talk- (2003-10)

A. Girardi, H. Singer, K. Shikano, and S. Nakamura, "Maximum Likelihood 
Successive State Splitting Algorithm for Tied-Mixture HMnet", IEICE 
transactions on information and systems, Vol.83 No.10, pp.1890-1897, 2000-10.

A. Girardi, K. Shikano, S. Nakamura, "Creating speaker independent HMM models 
for restricted database using STRAIGHT-TEMPO morphing", Proc. Internat. Conf. 
on Spoken Language Process, (1998-12) 

A. Girardi, H. Singer, K. Shikano, Satoshi Nakamura, "Maximum Likelihood 
Successive State Splitting Algorithm for Tied-Mixture HMNET", EUROSPEECH97, 
pp.119-122, 1997-9 

A. Girardi, H. Singer, K. Shikano, S. Nakamura, "Maximum likelihood successive 
state splitting algorithm for tied mixture HMnet", ASJ, 3-6-1, 89-90, 1997.3 

 
Employment Record:

Multitel ASBL - Belgium
 Period: Since Ago/2004 
 Position: Head of Speech Processing Adjunct
 Main Task: Management, research of new speech recognition and synthesis 
enhancement, development of speech dialog applications and coordination of new 
projects proposal and business models.

Federal University of Amazonas - UFAM - Brazil
 Period: Mar/2002 until Jul/2004
 Position: Master Course Teacher
 Main Task: Teach and orient master students (computer science, pattern 
recognition)

Siemens S.A. - Brazil
 Period: Dec/2003 until Jul/2004
 Position: Senior Consultant
 Main Task: Build the Manaus mobile phone R&D Center

European Commission - Europe
 Period: 2003 
 Position: Speech Science Expert
 Main Task: Evaluate applications for the IST FP6
 IST - Information Society Technologies
 FP6 - Sixth Framework Program

Genius Institute of Technology - Brazil
 Period: Dec/2001 until Oct/2003
 Position: ASR Program Leader
 Main Task: Business Development, Technology Guidance, MD Thesis advisor
 (PCA, ICA, LDA, SS, CMS, echo cancellation, FGT - Fast Gaussian Tree, 
 confidence measures, rejection, discriminative training, grammar optimization, 
 language modeling, microphone array, speaker identification, speech synthesis, 
 data validation)
 (version control, auto-documentation, auto-testing, UML, procedure 
 definitions, programming standardization, walk-through, total-quality, 
software 
 quality management, management by competence)
 (viability studies, business models, CMM implementation, government 
 relationship, funds submission, international programs, international 
 co-development, technology research and development co-ordination)

Lernout & Hauspie Speech Products S.A./N.V. - Belgium
 Period: Sep/1999 until Dec/2001
 Position: Research Scientist Senior
 Main Task: Research in Speech Recognition, Technology assessment
 (PCA, ICA, LDA, SS, CMS, echo cancellation, FGT - Fast Gaussian Tree, 
 confidence measures, rejection, discriminative training, grammar optimization)
 (Standardization boards relationship, international programs, international 
 co-development, technology research and development co-ordination)

ATR - Advanced Telecommunications and Research Laboratory - Japan
 Period: Mar/1996 until Mar/1997
 Position: Research Trainee (NAIST - ATR)
 Main Task: phonetic tree research
 (version control, auto-documentation, auto-testing, programming 
 standardization, total-quality, software quality management)

Chronos Sistemas Eletronicos S.A. - Brazil
 Period: Mar/1994 until Mar/1995
 Position: Development Manager
 Main Task: Manage product development (50 third part companies management), 
ABS 
 case development, production manufacturing and testing equipment development. 
 Risk Management, MS Project based tasks assignment and follow up. ISO 9001 
PDCA 
 (Plan Do Check Act) based management: including: brainstorm, Pareto, 
 responsibility tree.

Elo Sistemas Eletronicos S.A. - Brazil
 Period: Oct/1992 until Mar/1994
 Position: Development Supervisor
 Main Task: Supervise the development of embedded systems. I worked with FPGA 
 development, embedded microcontrolers, power management, fiber optics, 
 high-voltage, measurement class equipment development. 

Technological Institute of Aeronautics - ITA - Brazil
 Period: 1991
 Position: Graduate Teacher
 Main Task: Teach and orient graduate students (computer science, digital 
systems)

Elebra Comunicacao de Dados Ltda. - Brazil
 Period: Jan/1989 until Jul/1990
 Position: Design Engineer
 Main Task: Implement modem protocols, FPGA development, real-time OS, embedded 
 DSPs, coders, decoders and data compression

Computational Skills:
 Research: MATLAB, Octave
 Programming: (Voice)XML, C++/C, Python, Perl, Java (J2EE, J2ME)
 Legacy: Tcl-Tk, Pascal, Fortran, Basic and Cobol
 Assembler: ADSP 53X, C55, C54, OMAP, StrongARM, TMS320C25, 80x86, 80x51, etc...
 General: Office(97,2K) (Windows), Xemacs, OpenOffice, KDE (and applications)
 Engineering: AutoCAD, PCAD, (Quartus II, MaxPlus, APLUS - PLD - Altera).
 Profiling: VTune, Quantify, OptimalJava
 Environment: Win(95,97,NT,2K,XP) and Unix (RedHat 9, DEC OSF1 4.0D and SGI 
IRIX 6.3)
 Management: ClearCase (CVS), WinCVS, CPP, CDOC, MS Project, Primavera
 Testing: unitCPP, junit, Purify, Robot
 Modeling: Rational tools (RUP, Rose)
 Documentation:javadoc, doxigen 
 Database: SQL, MySQL
 Protocols: UPnP, WebServices, XML, VoiceXML 

Speech Recognition Related Software:
 HTK, Hark (from BBN), VoiceXpress (from L&H), C-Rec (from Dragon).

Recently, I coordinated the co-development of Speech2Go (together with 
SpeechWorks) and I have developed a proprietary version internally known at 
Genius as Dialeto (currently with improved performance compared to the previous 
mentioned softwares).
 
Management Tools (Multi cycle PDCA process):
 Plan: Brain storm and research (feasibility study, business plan - SWORT, 
CAPM), 
 schedule planning, SMART planning, test and benchmark planning, 
 documentation (product, API and procedures)
 Do: Document walkthrough, automatic documentation, time reports
 Check: (Measurement): automatic test reports, SMART review and feedback
 Act: Pareto

Psychology: Transactional analysis 
Methodologies: Risc management, competence management

Languages:

English: Fluent conversation, reading and writing very good.
Portuguese (BR): Fluent conversation, reading and writing very good (natural 
language).
Japanese: Fluent conversation, reading and writing good (4.5 years; Nagoya 
Univ., NAIST-Japan).
French: Good conversation, reading and writing good (Alliance Francaise).
Italian: Regular conversation, reading and writing good (1 year private course).
Spanish: Regular, reading and writing good.
Dutch: Basic level (2 years private course in Belgium)
German: Basic level (studying at moment).