Conf42 Machine Learning 2021
2021-07-29
-- online --
[Conf42 Machine Learning 2021, Conf42 Machine Learning 2022, Conf42 Machine Learning 2023, Conf42 Machine Learning 2025, Conf42 Machine Learning 2026]
who it's for
machine learning engineers, MLOps engineers, data scientists, ML platform engineers, DevOps engineers, security and compliance engineers, NLP engineers
topics
AWS SageMaker Pipelines; low-latency Python deployment; ONNX model export; Detectron2 and DETR object detection; transformers for object detection; vector database semantic search; serverless deep learning; model monitoring with Grafana; PyTorch convolutional networks; multilingual NLP pipelines; time series forecasting methods; HIPAA and regulatory ML controls
| title | author | length |
|---|---|---|
| Premiere | 13m | |
Pipeline automations show how to cut months of cross-team work to deployment by wiring AWS step-based ML pipelines into CI/CD; a drone case study benchmarks… | ||
| Automate your ML workflows with pipelines | Antje Barth | 45m |
Builds an end-to-end ML pipeline using Amazon SageMaker Pipelines that preprocesses text, generates embeddings, fine-tunes a pre-trained BERT model with a… | ||
| Responsible AI in Health | Tempest van Schaik | 20m |
A responsible AI review board evaluates projects with a checklist — is AI needed, who are stakeholders, what are benefits and harms, does data contain PII, and… | ||
| Convolutional Neural Networks in Action | Milecia McGregor | 24m |
Builds a small convolutional network in Python using PyTorch and tracks runs with DVC on the MNIST dataset, showing concrete layer choices: conv layers with… | ||
| Multilingual Natural Language Processing | Gajendra Deshpande | 23m |
Shows a concrete pipeline that extracts text from images, audio and web pages using pytesseract / PIL, SpeechRecognition and Beautiful Soup, detects language… | ||
| Security at your fingertips: Theory → Practice! | Madalina Burci | 24m |
Claims that keystroke dynamics work as a practical second factor: TypingDNA captures press/release and inter-key timings with an open-source recorder, creates… | ||
| Build ML environment for regulatory customers | Suraj Muraleedharan | 46m |
Regulated customers (banking, healthcare, pharma) get a hardened ML platform using AWS Service Catalog and SageMaker Studio: service catalog products shared… | ||
| Monitoring AI Pipelines Output As Product | Hila Fox | 25m |
Augury monitors ML pipeline outputs as product by instrumenting the detection lifecycle end-to-end: ~50 million machine-hours, ~80,000 machines, dozens of… | ||
| Object Detection using Transformers and CNNs | Eduardo Dixo | 28m |
A car detector is fine-tuned on a drone dataset of ~6,000 training and ~500 validation images, using Detectron2 and Hugging Face tools and Python code… | ||
| 10 Things That Can Go Wrong with ML Projects | Karl Weinmeister | 27m |
Lists ten concrete failure modes for ML projects and maps each to remedies: start with a prototype and baseline, avoid jumping straight to heavy models, expect… | ||
| Seamless Customer Service with Amazon Connect & Alexa | Asif Mujawar | 31m |
Demonstrates an Alexa-to-agent pipeline that uses Amazon Lex for NLP, Amazon Kendra for KB search, Lambda outbound calls, SageMaker personalization and Amazon… | ||
| Engineering Techniques for Binary IoT Sensors | Nidal Albeiruti | 32m |
Binary sensors produce more than on/off: the talk shows extracting elapsed-on duration, time-window aggregation, event frequency, inter-event delay and… | ||
| ML time series forecasting methods | Pawel Skrzypek, Anna Warno | 31m |
ES‑Hybrid pairs exponential smoothing with a dilated residual LSTM and wins M4; N‑Beats uses stacked blocks and ensembles 100+ models and claims better M4… | ||
| Leverage Power of Machine Learning with ONNX | Ron Lyle Dagdag | 41m |
Shows converting a small ML.NET linear regression trained on a 13‑row salary CSV into an ONNX file, then running inference from Python with ONNX Runtime to… | ||
| Search through your data with Weaviate | Laura Ham | 13m |
Live demo shows Weaviate ingesting a news corpus, vectorizing text (and images) with built‑in models, and serving vectorized semantic search plus filters via… | ||
| Deploying ML solutions with low latency in Python | Aditya Lohia | 18m |
Shows concrete steps and numbers for cutting inference latency: convert PyTorch/TensorFlow models to ONNX then optimize with TensorRT using FP16/INT8… | ||
| Empower your business use cases | Matteo Gabrielli | 26m |
A working Document Understanding Solution assembles Amazon Textract for OCR and table-to-CSV extraction, Amazon Comprehend Medical for medical-entity detection… | ||
| Serverless Deep Learning | Nicola Pietroluongo | 12m |
Demonstrates running inference as a serverless API by bundling a PyTorch MNIST model into AWS Lambda container images, building and testing locally with the… | ||
| E2E ML Platform on Kubernetes with a few clicks | Mofi Rahman | 30m |
Demonstrates getting a full Kubeflow E2E platform up on Kubernetes using the Kubeflow Operator to automate deploy, monitoring and lifecycle management… | ||
| Natural language modelling- BlazingText algorithm | Dinesh Subramani | 47m |
Demonstrates producing word embeddings from a Tamil Wikipedia dump using the BlazingText algorithm packaged in Amazon SageMaker, with a distributed training… | ||
| Deploying ML Models and the Things That Go Wrong | Marianna Diachuk | 29m |
Hands-on account of deploying models since 2017 with concrete tradeoffs: choosing dedicated deployment engineers versus expecting data scientists to own… | ||
| Unveiling the secrets of ancient coins | Nicolas Metallo | 37m |
University of Oxford’s 21 million‑object backlog becomes a quantifiable ML prototype: a heads vs tails classifier trained via Amazon SageMaker (JumpStart… | ||
| Hail Hydrate! From Stream to Lake | Timothy J Spann | 36m |
Claims a runnable FLaNK pipeline that uses Apache NiFi, Pulsar, Flink, MiNiFi to ingest IoT and NOAA weather feeds (NOAA ZIPs every ~15 minutes), convert… | ||
| Pragmatic Machine Learning in the Cloud | Joshua Arvin Lat | 44m |
Shows concrete patterns and working code to run ML on AWS SageMaker: built-in algorithms (about 17 algorithms), dataset processing, Debugger snapshots to S3… | ||
| nothing matches | ||