🥇Top-tier product-first AI consultant.
( Expert-Vetted=Upwork verified Data Expert Top 1%)
----
💁🏻Data engineer, business minded Data scientist, Ex software engineer.
6+ years, 30+ end-to-end data science and machine learning projects delivered with 5⭐ reviews.
🤝🏻I specialize in Natural language Processing (From NLTK - Transformers - Large language models like Claude, Mistral, GPT-4 ), Computer vision and Time-series analysis.
With solid base in Data Engineering and MLOps, I have engineered and deployed cutting-edge tech in LLMs and NLP, Computer vision and Data Science.
I enjoy building in automated ML workflows with active learning i.e., continuous model life cycle monitoring and improvement in model performance metrics by new data streaming.
🎯Projects:
🚩Product development - Scalable social media brand intelligence product with real time data. E2E, data engineering to MLops, LLMs integrations. Product similar to Brandwatch.
📈MVP: Time series forecasting for a tech company in water treatment, based in NL, complete data pipeline construction, feedback-driven continuously improving time-series ML / DL models.
🛠Prototype development: Integration of Computer Vision using DL, automate wide range of tasks in the manufacturing such as product classification, component localization, defect detection, object counting, segmentation, OCR.
🚩Product development - Leveraging NLP and Data science to build more inclusive + productive team management for an Australian startup with in-depth *anomalies* detection.
⚡Feature enhancements - Combining vision and text-to-speech to build intuitive branding videos - a swiss startup.
🏩Prototype development - NL based interior decoration startup required a computer vision app to visualize new décor concepts.
🔄NLP pipeline - For an Italian startup, planned and built entire pipeline, with active learning NLP models i.e. continuously improving with incoming data. Seamless Integration of the Transformers for various use cases such as Named Entity Recognition (NER), and Gen AI apps.
💪Prototype development - For an NY based fitness startup, built a SOTA human pose estimation model - lite weight mobile deployed models with optimal accuracy and response times.
🙋Product development - built transformer based highly engaging assistant for a mental health startup based in Germany.
🕵New feature development - To detect hacking threats with NLP analysis for a gaming industry startup in NL.
📊POC - stock price prediction - (forecast, nowcast) using Echo State Networks (ESN).
💊New feature development - Custom Spacy NLP pipelines in bio medical and clinical domain for a Bengaluru based startup.
Publications:
🧠 "Machine Learning Methods for Managing Parkinson’s Disease" published by Springer
🍅 "Using Deep learning for image based plant disease detection" published by De Gruyter
🌻 "Internet of things platform for smart farming" published by De Gruyter
Services:
Data mining
EDA
Data pipeline strategizing
Upgrade existing products and services with Machine learning (ML), Deep learning (DL)
Data-driven Decision support systems,
Startup product development
Hands on:
Classification, Regression, Neural Networks ( CNN, RNN, bi-directional LSTM), Recommender Systems, Support Vector Machines (SVM), K-Means and Principal Component Analysis (PCA), Reinforcement Learning, Natural Language Processing (NLP), Computer vision( 2D/3D classification, detection, segmentation, pose estimation)
DL libraries:
TensorFlow,PyTorch, MACE, MXNet, fastAI, tflite
Databases:
MySQL, Postgres, MongoDB, Neo4js, Elastic Search
Data wrangling, text mining, data analysis and data visualization Tools:
Pandas, NumPy, Scikit-learn, Matplotlib, Seaborn, Plotly
Data pipelines - Active learning:
Prodigy, Label Studio, Label box, AWS, GCP
NLP :
Data scrapping, Parsing, Annotation, POS tagging, Stemming, Tokenization,
Named entity extraction, Text classification, Machine translation, Text summarization, Q/A, Text generation, Sentiment analysis, Topic segmentation, chatbots
With,
Word2vec, Doc2vec, NLTK, Gensim, Scispacy, SpaCy 3, Dialogflow, RASA, BERT, huggingface
LLMs: LLAMA 2, Claude, GPT 4, Mistral Langchain
Computer vision :
Image Classification, Object Detection & counting, Image Segmentation, Human pose estimation
With,
OpenCV, MobileNet, Efficientnet, Inception, DarkNet , Faster R-CNN, UNet, SegNet, Mask R-CNN, DeepLab, TF, Pytorch, MXNet, YOLO, COCO, Openpose, HR net, Posenet, Mediapipe, Detectron
MLops:
MLflow, AWS sagemaker, GCP
Product development / MVP/ POC:
Streamlit, Flask, Fast API
Special interest in:Healthcare, Fintech, Climate, Green Energy, Edtech