Farhad Imani
Sr Staff Applied Ml Scientist Team Lead @ThoughtExchange
Signup · Get unlimited contacts
WORK HISTORY
Sr Staff Applied Ml Scientist Team Lead @ThoughtExchange
CA
Thoughtexchange is a new way to lead challenging conversations about things that matter. Thoughtexchange provides software and professional services to let you lead productive group conversations with 15 to people. a team lead, I have been responsible for leading the ML team, and directing the ML roadmap. As an individual contributor, I have been in charge of the design, development and implementation of Machine Learning (ML) and Natural Language Processing (NLP) algorithms into product. Multiple tools, platforms and libraries were used including LSA, LDA, NMF, Word2Vec, Doc2Vec, page ranking, CNN, LSTM, combination of CNN and LSTM, BERT, DISTILBERT, T5, GPT-3, GPT-4, RAG, different clustering techniques and several Python libraries using CPU and AWS GPU instances. Achievements:• Led the ML team in collaboration with the dev team and contributed to the R&D roadmap.• Supervised multiple team members.• Investigated, designed and integrated LLM-based solutions including the data preparation, fine-tuning, in-house RAG, agentic RAG, prompt engineering and evaluation using GPT and Anthropic models.• Designed, developed and implemented a fully automatic unsupervised approach to categorize similar thoughts using combination of topic modeling, clustering and topic stability techniques.• Designed, developed and implemented the search and tagging solutions using semantic and keywords hybrids algorithms by applying the BERT and DISTILBERT models fine-tuned for text similarity.• Designed, developed and implemented deep learning networks using CNN, LSTM and their combination to flag and moderate pieces of text using various criteria.• Designed, developed and implemented a ranking algorithm to rank information using the textual and rating data by applying the embedding of words and graph creation.• Designed, developed and implemented an algorithm to detect near duplicate information using the textual and rating data.
EDUCATION
Queen's University
Doctor of Philosophy (PhD), Electrical Engineering
Iran University of Science and Technology
Bachelor of Applied Science (B.A.Sc.), Electrical and Electronics Engineering
Iran University of Science and Technology
Master of Science (MSc), Electrical Engineering (Bio-Electrics)
SKILLS
ABOUT FARHAD IMANI
I am a Machine Learning specialist with extensive expertise in signal, image, and natural language processing. My career spans significant contributions in developing advanced algorithms and frameworks across diverse domains including medical image analysis, text mining, and video content optimization.With a robust background that includes roles at Thoughtexchange, BroadbandTV, and collaborative projects with institutions like Philips and NIH, I have led teams and driven research in applying cutting-edge technologies.I am passionate about leveraging my skills in developing innovative software systems tailored for signal, image, and NLP applications. I am eager to join an organization that fosters growth and innovation in these fields. Skills- Programming Fundamentals: Data Structure, Algorithm Design, Object-Oriented Programming, Scrum-Based Agile Development- Programming Languages: Python, C/C++, SQL, Lua, Assembly- Tools & Platforms: Keras, TensorFlow, Torch, Redis, OpenCL, OpenCV, NLTK, Scikit-Learn, MATLAB, Simulink, Multiprocessing, Threading- Machine Learning Approaches: Clustering, Principal component analysis (PCA), Independent Component Analysis (ICA), joint Independent Component Analysis (jICA), t-distributed Stochastic Neighbor Embedding (t-SNE), Recursive Feature Elimination (RFE), Support Vector Machine (SVM), Random Forest, Multilayer Perceptron (MLP), Deep Belief Network (DBN), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Latent Semantic Analysis (LSA), Latent Dirichlet Allocation (LDA), None-negative Matrix Factorization (NMF), Word2Vec, Global Vectors for Word Representation (GloVe), Transformer-based models including Bidirectional Encoder Representations from Transformers (BERT), DISTILBERT, T5, GPT-3 and GPT-4, Anthropic Claude, Fine-tuning, RAG, LlamaIndex.Thanks for viewing my profile. Please scroll down to see a complete and detailed list of my accomplishments and project\'s achievements. If you\'re interested in starting a conversation, I can be reached on 60••••••74 or by email at: f••••••••@gmail.com
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.