Ivan Podluzhnyi, Applied Scientist
| Cambridge, GB
SUMMARY
I'm a full stack web developer who can build apps from the ground up. I've worked mostly at startups so I am used to wearing many hats. I am a very product focussed developer who priotizes user feedback first and foremost. I'm generally very flexible when investigating new roles.
EDUCATION
Saint Petersburg State University 2015 — 2020
Masters - Probability Theory and Statistics
ITMO University 2024 — 2020
Ph.D. (incomplete) - Computer Science
SKILLS
Areas of expertise (Advanced): Automatic Speech Recognition, Natural Language Processing (NLP), Speech Signal Processing, Article writing, Data Science
Areas of expertise (Intermediate): Computer Vision, Machine Learning
Frameworks (Advanced): Pytorch, Espnet, NeMo, Kaldi
EXPERIENCE
| Applied Scientist 2023-05 — Present

TTS

| Speech Research Scientist 2019-08 — 2022-11
  • Created and released hybrid speech recognition models for different languages (Russian,English, Spanish, Kazakh, Turkish, Arabic) with 80%+ accuracy
  • Developed BPE-dropout pipeline in ASR training, increasing OOV recognition rate by 25%
  • Developed and released NLP punctuation model for ASR results with 0.8 F1 score
PUBLICATIONS
LT-LM: A Novel Non-Autoregressive Language Model for Single-Shot Lattice Rescoring Aug 30, 2021
Proc. Interspeech 2021
Dynamic Acoustic Unit Augmentation with BPE-Dropout for Low-Resource End-to-End Speech Recognition Apr 28, 2021
MDPI Sensors
Towards a Competitive End-to-End Speech Recognition for CHiME-6 Dinner Party Transcription Oct 26, 2020
Proc. Interspeech 2020
Target-Speaker Voice Activity Detection: A Novel Approach for Multi-Speaker Diarization in a Dinner Party Scenario Oct 25, 2020
Proc. Interspeech 2020
Three properties of a discrete dynamical system in the space of infinitely differentiable functions Jan 2019
Differencialnie Uravnenia i Protsesy Upravlenia, 2019
AWARDS
CHiME-6 challenge Track 2 winner 2020-05-31
CHiME-6 challenge committee

As a part of the joint STC-innovations Ltd & ITMO University team, won Track 2 of the CHiME-6 challenge. More at https://chimechallenge.github.io/chime6/results.html

LANGUAGES
ru (Native Speaker) , en (Advanced)