About
Technical Lead with team leadership, architecture, and C++/C#/Python Application and…
Services
Activity
11K followers
-
eric dransfeldt shared thiseric dransfeldt shared this🌟 Exciting Update for the Deep Signal Library! 🌟 Thrilled to unveil some powerful features in Version 1.6 that are set to amplify your trading game! 🚀 Here's the scoop on what's new: 1️⃣ Permutation Feature Importance: Dive deep into your models! Now, with Feature Importance, you can scrutinize the impact of added features on your models. Uncover the gems that truly contribute to model excellence. 🕵️♂️✨ 2️⃣ Dataset Weights for Precision: Achieve balance effortlessly! Introducing Dataset Weights, where you can assign weights to Profit Target and Profit Target Not Reached datasets. Whether you prefer automatic balance or want to customize weights, the power is in your hands! ⚖️💰 3️⃣ Trainer Customization Bonanza: Elevate your model creation process with a plethora of customization options! Explore the new Scda, Lbfgs, Fast Forest, Fast Tree, and Lgbm Trainer Options. Fine-tune to perfection with control over L1/L2 Regularization, Leaves, Trees, Feature Fraction, Example Count, Bin Count, and Learning Rate. 🛠️🌲 #deepsignaltech #ninjatrader #machinelearning #tradingstrategy
-
eric dransfeldt shared thiseric dransfeldt shared thisDeep Signal is excited to introduce our Training Data Viewer for the Deep Signal Library that allows the trader to visualize what data is being used to train their machine learning models. The chart will display where the Pre-Signal window, Signal Bar and Bars To Target window are located in relation to the data that was used to create a model. For more information on how to use the Training Data Viewer with the Deep Signal Library, please see our Online Help. #deepsignaltech #ninjatrader #machinelearning #tradingstrategy
-
eric dransfeldt shared thiseric dransfeldt shared thisDeep Signal is excited to announce the release of the Deep Signal Library 1.5. The new version includes Regression and Multiclass machine learning trainers for creating new financial trading models. Please download the latest Deep Signal Library and try out the new trainers. #NinjaTrader #TradingStrategy #DeepSignalTech
-
eric dransfeldt shared thiseric dransfeldt shared thisWant to use multiple machine learning models for trading different market conditions or triggers to enter a trade? The Deep Signal Library can now use multiple models in one strategy. Please download the latest Deep Signal Library to try it out! #NinjaTrader #TradingStrategy #DeepSignalTech
-
eric dransfeldt shared thisGreat list of courses! Thank you for sharing!eric dransfeldt shared thisSo you are looking for Machine Learning courses on Youtube? Sure, here you go! You are welcome! - Andrew Ng CS229 ML: https://lnkd.in/gkDEyuCS - MIT: Deep Learning for Art: https://lnkd.in/grusgt3Z - Stanford CS230: Deep Learning: https://lnkd.in/ggXNEX7K - Practical Deep Learning for Coders: https://lnkd.in/giHMNrHG, https://lnkd.in/gDtRtHmG - Stanford CS224W: Machine Learning with Graphs: https://lnkd.in/grZC_j4N - Probabilistic Machine Learning: https://lnkd.in/gjSpNDCD - MIT 6.S191: Introduction to Deep Learning: https://lnkd.in/gWtSdkSH - UC Berkeley CS 182: Deep Learning: https://lnkd.in/gzHS6m8G - UC Berkeley Deep Unsupervised Learning: https://lnkd.in/gPdPbKku - Yann Lecun's NYU Deep Learning SP21L: https://lnkd.in/gdyzmf8b - Stanford CS25 - Transformers United: https://lnkd.in/gaZVn3wY - Hugging Face NLP Course: https://lnkd.in/gigfE2Yj - Stanford CS224N: Natural Language Processing with Deep Learning: https://lnkd.in/g4fg4_wX - CMU Neural Nets for NLP: https://lnkd.in/gVpUwtXE - Stanford CS224U: Natural Language Understanding: https://lnkd.in/gMeGkkzV - CMU Advanced NLP: https://lnkd.in/gAtrsGqY - CMU Multilingual NLP: https://lnkd.in/ghbcWftV - Stanford CS231N: Convolutional Neural Networks for Visual Recognition: https://lnkd.in/g3DeCWEc - Michigan Deep Learning for Computer Vision: https://lnkd.in/gbdgGgJQ - AMMI Geometric Deep Learning Course: https://lnkd.in/gYH6Vuum - UC Berkeley CS 285 Deep Reinforcement Learning: https://lnkd.in/gH-HYdqz - Intro to Deep Learning and Generative Models: https://lnkd.in/gxuTtkSk - Stanford CS330: Deep Multi-Task and Meta Learning: https://lnkd.in/gasntdBh Source: https://lnkd.in/gys5Rk5k Looking for career mentoring and/or Machine Learning consulting services? Let's chat: https://lnkd.in/gGBMXuR4 #machinelearning #deeplearning
-
eric dransfeldt shared thiseric dransfeldt shared thisWe're excited to announce the Deep Signal Library version 1.3 has been released. It includes the ability to choose what trainer to use when creating a machine learning model. You can still have the Deep Signal Library try all available trainers when creating a model to find the best performing algorithm. The progress window that is displayed when creating a new machine learning model has been updated. It shows the best performing trainer in real time as the model is being created. It also will keep track of each trainer's performance so the user can look to see how each trainer performed during the training run. If you find some trainers do not work well for that dataset then you can deselect them for future training runs. #deepsignaltech #ninjatrader #tradingstrategy
-
eric dransfeldt shared thiseric dransfeldt shared thisThis eBook takes a deep dive into many common software architecture patterns. Free, our gift to you.
-
eric dransfeldt shared thiseric dransfeldt shared thisWe are excited to announce that the Deep Signal Machine Learning Library for NinjaTrader is available for download. After two years of development and a successful beta program, the Deep Signal Library is available for traders who want to create and use machine learning models for financial trading. The library is an extension of NinjaTrader that automates the process of creating a machine learning model that can be used in trading. #machinelearning #trading #ninjatrader #deepsignaltech
-
eric dransfeldt shared thiseric dransfeldt shared thisWe are hiring research interns at Microsoft Mixed Reality! If you are a PhD student interested in deep learning and computer vision, and want to see how research gets turned into world-class devices like the #HoloLens, consider submitting your resume at the following link: https://lnkd.in/ghJngRM #microsoft #hiring #computervision #deeplearning
-
eric dransfeldt liked thiseric dransfeldt liked thisOver 80% of published investment factors are likely false. In a new paper, Frank Fabozzi and I argue that the core problem is not only multiple testing. It is identification failure. Our work reconciles the apparent contradiction between academic claims and investors' experience: it answers why academic studies estimate that the False Discovery Rate (FDR) in finance is only 10%, while investors experience that most factor investment funds fail to perform. The reason is, published statistics are not outcomes of a single trial. They are the result of selecting the most favorable outcome among many within-study candidate specifications. That changes the statistical experiment. And once the experiment changes, standard inference breaks. Main messages: - The FDR cannot be identified from published in-sample statistics alone - Under standard selection rules, single-trial inference understates the true FDR - Very different underlying worlds can generate very similar reported significance profiles - Under an explicit search-adjusted parametric model, the fitted FDR exceeds 80% Practical implication for academics, practitioners, and investors: Statistical significance is not enough Reliable inference requires either: - an explicit model of the search-and-selection process, or - genuinely independent validation data A broader lesson is that complexity is not a virtue when it merely hides specification search. A more reliable path is causal factor investing: signals grounded in economically coherent mechanisms, not just in-sample winners. Paper: https://lnkd.in/d96CEN8z Data & Code: https://lnkd.in/drWephNi #Finance #AssetPricing #FactorInvesting #Quant #MachineLearning #Statistics #FalseDiscoveryRate #BacktestOverfitting #CausalInference #ResearchGitHub - lopezdeprado/FDR-in-Finance: Code for replicating the paper "The False Discovery Rate in Finance: Identification Failure and Search-Adjusted Estimation".GitHub - lopezdeprado/FDR-in-Finance: Code for replicating the paper "The False Discovery Rate in Finance: Identification Failure and Search-Adjusted Estimation".
-
eric dransfeldt liked thiseric dransfeldt liked thisMost teams deploy LLMs with default settings and wonder why inference costs $50K/month. The optimization stack exists. Most engineers don't know the layers. Here's the full inference optimization hierarchy: LAYER 1: Serving architecture Before touching a single kernel, get your serving right. vLLM (74K ⭐): PagedAttention, continuous batching. https://lnkd.in/eeT_HM2B SGLang (25K ⭐): structured generation + RadixAttention. Faster for constrained outputs. https://lnkd.in/eKK7sxdf LAYER 2: Quantization Shrink the model without killing accuracy. llama.cpp (92K ⭐): GGUF quantization. Run 70B on consumer hardware. https://lnkd.in/eJrUg_qd Unsloth (50K ⭐): QLoRA fine-tuning at 70% less VRAM. https://lnkd.in/gJZtH4Y4 This layer alone can cut your GPU bill in half. LAYER 3: Attention + caching How much are you spending on redundant prefill? Flash Attention (21K ⭐): memory-efficient, IO-aware. Non-negotiable. https://lnkd.in/eYkuRuxC LMCache (1.5K ⭐): KV cache sharing. Eliminates it entirely. github.com/LMCache/LMCache LAYER 4: Hardware-specific acceleration Match your optimization to your silicon. TensorRT-LLM: purpose-built for NVIDIA GPUs. Kernel fusion, in-flight batching. https://lnkd.in/ekuFuDAP MLX: native framework for Apple Silicon. Inference without CUDA. github.com/ml-explore/mlx LAYER 5: Custom kernels Where the real differentiation lives. LeetCUDA (9K ⭐): 200+ CUDA kernels. Tensor Cores, HGEMM. https://lnkd.in/eUfgpwW6 llm.c (28K ⭐): Karpathy's raw C/CUDA. The fundamentals. github.com/karpathy/llm.c Engineers who write custom kernels command $200K+ at NVIDIA, Meta, and Google. LAYER 6: Distributed inference When one node isn't enough. NVIDIA Dynamo: multi-node orchestration. Disaggregated serving. https://lnkd.in/etBGNtjk exo (39K ⭐): distributed inference across consumer devices. github.com/exo-explore/exo 6 layers. Each one multiplies the savings from the layer above. Most teams stop at Layer 1. The ones running inference profitably reach Layer 5. Which layer is your team stuck at? 👇 💾 Bookmark this. Your next inference bill will thank you.
-
eric dransfeldt liked thiseric dransfeldt liked thisi'm starting a newsletter called Owners, Not Renters - because closed AI isn't winning because it's better. it's winning because it's easier. we've been here before. a dominant platform, everyone locked in, and then, eventually, a shift. open becomes the practical choice, not just the principled one. that shift is starting! 70b models on a mac. ollama pull and you're running. developers swapping out cursor, swapping out claude code, for opencode and local weights and local context. your data never leaves your machine. the defaults are hardening. that's the part that keeps me up. i want to see wonderful things built on open. weird, surprising, new experiences we can't imagine yet. but you don't get there until you get the tools right. maslow's hierarchy. before wonderful comes: how do i actually get this running? i don't know yet. i'm trying to figure it out. i'll tell you what i'm learning - what models are actually good, where the stack is still duct tape, what's working and what isn't. and i want to hear what you're learning. subscribe. reply to the first one. i'm reading everything. link in comments.
-
eric dransfeldt liked thiseric dransfeldt liked thisAn open-source alternative to Claude! (18k+ stars; works with any LLM) Onyx is a self-hostable AI chat platform that works with any LLM, like Claude, GPT, Gemini, Llama, or any open-weight model you want. Here's what it ships with: - Agents that chain multiple tools in sequence - RAG with full indexing across 40+ connectors (Slack, Drive, Confluence, Jira, GitHub, email, call transcripts) - Deep research ranked #1 on DeepResearchBench, above every proprietary alternative - MCP support for connecting to external systems - Code interpreter for data analysis and file generation - Self-host on your own infrastructure via Docker in a few mins Unlike Claude's MCP-based connectors that query your tools at runtime, Onyx actually indexes all your data. That means faster, more reliable search across everything your team has ever written. The entire code is open-source under MIT license, so you can see the full implementation on GitHub and try it yourself. Link to the repo in the first comment! _____ Share this with your network if you found this insightful ♻️ Follow me (Akshay Pachaar) for more insights and tutorials on AI and Machine Learning!
-
eric dransfeldt liked thiseric dransfeldt liked thisAnthropic is coming to Portland, yes THAT Anthropic... pt.2! Our first Claude event hit 500+ RSVPs and sold out in HOURS - so we're running it back for pt.2! If you didn't get a chance to be in the room the first time, here is your chance now. We're hosting a Claude meetup to show the world what we're capable of. This is our chance to put Portland on the map + prove that we DO belong in the AI conversation. None of this happens without this incredible group of people - massive shoutout to Dinesh Mathew, Abraham N., Todd Greco, Dave Barcos, Sam Keen, Joanna Gough, Natalie Tashchuk, Grant Macdonald, Dominic Kuklawood, Rick Turoczy, Mike Biglan, M.S. + many more. These people showed up when it mattered and are bringing our communities together in the AI era. Now it's your turn. Seats are extremely limited. This will sell out. Period. Comment "Claude" and I'll send you the link to apply. (Make sure we're connected)
-
eric dransfeldt liked thiseric dransfeldt liked thisIt was an honor to hang out with Jensen Huang, CEO of NVIDIA, and do a long-form podcast with him. Really fun & fascinating technical deep-dive conversation on & off the mic. One of the most brilliant & thoughtful human beings I've ever met. NVIDIA is the most valuable company in the world by market cap and is the engine powering the AI revolution. Podcast probably out tomorrow (Monday), unless I get stuck in too many interesting conversations while running around in SF ;-) PS: I haven't checked my messages in days. Sorry for slow replies 🙏 Trying to stay deeply focused in an overwhelmingly intense time & barely hanging on. Love you all! ❤️
Experience & Education
-
Deltek + ComputerEase
View eric’s full experience
See their title, tenure and more.
or
Patents
-
Method and apparatus for controlling temperature response of a part in a conveyorized thermal processor
US 6,453,219
Other inventors -
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content