Μηχανική Μάθηση (ML) Jobs

We're looking for engineers and other STEM professionals to help push the limits of AI reasoning. Each task is simple in shape but hard in substance: you take a real image from your area of expertise and pair it with a challenging problem that a non-expert couldn't solve. What you'd do: - Choose or take a real, copyright-clear image with no annotations or markups - Write a difficult question about it that draws on your professional knowledge - Provide a clear, correct answer and reasoning Who this is for: - Working engineers (mechanical, civil, electrical, structural, software, etc.) and other STEM professionals - People who can spot the details in an image that only an expert would notice - Strong written English and attention to detail

I have a concept that relies on automatically classifying blocks of text, and I’m ready to move from idea to a production-ready machine learning model. The data is already in hand—clean, labeled text that spans several categories—and I want a purpose-built classification solution trained on this set. Here’s what I need from you: • Prepare and explore the text data, handling any necessary preprocessing or feature engineering (tokenisation, vectorisation, etc.). • Design, train, and tune a machine learning architecture suited to text classification. I’m open to traditional approaches (e.g., scikit-learn, XGBoost) or neural models (TensorFlow / PyTorch) as long as performance and interpretability remain strong. • Evaluate the model with clear...

I'm looking for an experienced AI trainer to conduct a course on Computer Vision. The training will be for intermediate learners and will cover the following topics: - Image Classification - Object Detection - Image Segmentation Ideal Skills and Experience: - Proficiency in Computer Vision techniques and tools - Experience in teaching or training, especially at an intermediate level - Strong understanding of Image Classification, Object Detection, and Image Segmentation - Ability to create structured, engaging, and informative training materials - Patience and good communication skills Please provide a portfolio or examples of previous training conducted. Looking forward to your bids!

I’m examining how Rwanda’s pioneering drone-delivery network can move even faster by sharpening the delivery-algorithm layer. My sole focus is on the technological side—specifically the mathematics, data structures and decision logic that pick, sequence and adjust each flight path in real time. What I need from you is a compact but rigorous study that shows, in measurable terms, how alternative or refined algorithms could cut overall delivery time. Feel free to draw on graph-search techniques, heuristic or meta-heuristic optimisation, predictive analytics, or machine-learning-supported routing—whatever you judge most promising for Rwanda’s geography, air-corridor rules and demand patterns. Deliverables • Literature and market review of algorithms curr...

We are seeking an experienced AI/ML Engineer to upgrade an existing AI-powered Fraud Detection System. The role focuses on improving fraud detection accuracy, real-time risk scoring, anomaly detection, and automated investigation. Responsibilities - Audit and upgrade the existing fraud detection architecture. - Improve AI/ML models for fraud detection, risk scoring, and anomaly detection. - Develop AI agents to investigate suspicious transactions and identify fraud patterns. - Integrate transaction, customer, device, behavioral, and external risk data. - Reduce false positives while detecting emerging fraud patterns. - Implement real-time monitoring, alerts, explainable risk scores, and investigation summaries. - Build testing, evaluation, monitoring, and continuous improvement processes....

Interview Engineer We’re looking for an Interview Engineer who is confident communicating in English and has strong Python skills from LATAM. What You’ll Do Conduct and facilitate technical interviews, particularly for engineering and AI-related roles. Evaluate candidates’ technical skills, problem-solving abilities, and communication. Design and improve technical interview questions and coding assessments. Work with the hiring team to provide clear and structured candidate feedback. Stay up to date with modern software engineering and AI technologies. Requirements Strong English communication skills, both written and spoken. Strong Python programming skills. Solid understanding of software engineering fundamentals. Excellent problem-solving and analytical skills. Abili...

Interview Engineer We’re looking for an Interview Engineer who is confident communicating in English and has strong Python skills from LATAM. What You’ll Do Conduct and facilitate technical interviews, particularly for engineering and AI-related roles. Evaluate candidates’ technical skills, problem-solving abilities, and communication. Design and improve technical interview questions and coding assessments. Work with the hiring team to provide clear and structured candidate feedback. Stay up to date with modern software engineering and AI technologies. Requirements Strong English communication skills, both written and spoken. Strong Python programming skills. Solid understanding of software engineering fundamentals. Excellent problem-solving and analytical skills. Abili...

I run a custom “API runner” that chains together multiple LLMs for internal red team research, and I need a seasoned LLM red-teamer to test it with various prompt engineering methods to see where its weaknesses are. Must have experience with JB and prompt engineering with proven success. Deliverable: • JB prompt that reveals model weakness. Skills: - Red Team - Prompt Engineering - PAIR and TAP - Pentesting - JB - English speaking - Available in PT timezone Let’s make this runner both smarter and safer.

Senior Full-Stack Developer — Production-Ready Data Intelligence Platform I am looking for an experienced senior full-stack/platform developer to take an existing data intelligence platform from its current development state to a secure, stable and production-ready product. This is not a greenfield build. A substantial application, architecture, database, frontend/backend and technical documentation already exist. The objective is to review what has been built against the intended architecture, identify gaps and technical debt, and complete the work required for production deployment. Platform Overview The platform is an enterprise data intelligence application that integrates data from customer business systems, processes and standardises it into a canonical data model, and prov...

I have Arabic Multimodal Dataset (Text + Image) designed for Multimodal Aspect-Based Sentiment Analysis (MABSA). Each data point consists of an Arabic review text paired with one or more relevant images. I am seeking an expert LLM Specialist to build a comprehensive, state-of-the-art experimental framework to validate the dataset's learnability and establish standard benchmarks for the scientific community. The project involves task-specific fine-tuning using modern text and vision-language architectures The dataset must be evaluated across two distinct tasks following standard evaluation benchmarks (e.g., SemEval standards): Task 1: Aspect Category Detection (Identifying specific predefined aspect categories mentioned). Task 2: Aspect-Based Sentiment Polarity Classification (Classify...

I’m setting up a weekday programme for absolute beginners who want to break into Artificial Intelligence,Machine Learning, MERN Full Stack, DevOps . You will run live, interactive classes for two hours each day, Monday through Friday, with the primary goal of giving students a rock-solid grounding in AI/ML basics—especially supervised learning. Format & schedule • 2 hr live session, five days a week (Mon–Fri) • Start date and exact time slot to be agreed together • Remote delivery via Zoom, Google Meet or a comparable platform Core content Students must leave the course confident in the fundamentals: data preparation, model selection, training & evaluation, and how supervised learning differs from other paradigms. Feel free to weave in...

Sensei AI — Institutional-Grade Stock Analysis Platform I built Sensei AI, a real-time trading intelligence platform covering all 50 Nifty stocks, designed to bring institutional-level AI analysis to retail investors. The system unifies five distinct AI model families into a single scored decision engine (−5.0 to +5.0): - Classical ML — Random Forest with SHAP-based feature attribution for explainable UP/DOWN probability - Deep Learning — Custom LSTM and causal Temporal CNN (PyTorch) for 5-day return forecasting - Reinforcement Learning — PPO agent (Stable-Baselines3) trained via a custom Gymnasium trading environment for BUY/SELL/HOLD actions - Regime Detection — Gaussian Hidden Markov Model to classify BULL/BEAR market states - Financia...

I am looking for an experienced AI/ML developer to build and integrate an advanced AI video transformation feature into my existing portal. The system should create realistic face movements and body motions to simulate live yoga sessions. I will provide a source video, my visual references, and voice/audio samples. The aim is to replace the source face with mine while preserving natural expressions, head movement, body gestures during yoga poses, as well as lip synchronization with the provided voice. It should convincingly mimic a live session's feel. Applicants should have expertise in AI video generation, face replacement, body motion tracking, voice cloning, and backend integration techniques. I'm open to suggestions on using suitable AI models, APIs, or a self-hosted soluti...

I need help turning a raw idea for group development into a practical, engaging series of team-building workshops for a cohort of 10–20 participants. The job begins with shaping a clear learning arc, then moves into designing hands-on activities that strengthen collaboration, communication and trust. I already know the workshops must be delivered in a group setting, so every exercise has to be interactive—think problem-solving challenges, role-play scenarios, and reflection debriefs rather than slide-heavy lectures. Here is what I will count on you to provide and do: • A concise workshop blueprint (timed agenda, objectives, and flow) ready for either an in-person room or a virtual platform such as Zoom with breakout rooms and Miro boards. • All participant material...

I own Celero Digital LLC, an Arizona IT and software consulting company. I am looking for a U.S.-based sales and lead-generation professional who can help me find real B2B IT/software clients. My main services include .NET/C#, ASP.NET, Azure, AWS, cloud development, DevOps, Kubernetes, microservices, AI/ML, cybersecurity, and healthcare software. My budget is a $450 one-time payment for client acquisition and sales outreach. I do not want a monthly retainer. I need the person to: * Find qualified U.S. companies that need IT/software consulting services. * Contact decision makers by email, phone, LinkedIn, or other appropriate methods. * Introduce Celero Digital and our services. * Generate qualified leads and schedule meetings with interested companies. * Follow up with prospects. * He...

Social Media Manager for Auto Detailing Business — $100/Month — Long-Term I own Porter’s Detailing, an auto detailing, paint correction, ceramic coating, and window tint business in Illinois. I’m looking for ONE reliable person to handle my social media on an ongoing basis. I’m not looking for a large marketing agency. I want to build a long-term relationship with one freelancer who understands the automotive/detailing industry and can make our social media look professional and active. Budget: $100/month fixed I will provide you with the raw photos and videos from the vehicles we work on. You will NOT need to travel, photograph vehicles, or create content from scratch. What I Need * Turn my raw photos/videos into professional social media content * Crea...

Looking for an experienced developer to build an AI-based UAV detection and tracking system for a custom ArduPilot quadcopter using Raspberry Pi 5 and AI HAT+ 2. Scope includes real-time drone detection, object tracking, MAVLink integration, safe autonomous follow/stand-off navigation, telemetry logging, GCS interface, failsafes, SITL testing, and complete source-code handover.

I have a single video frame captured in daylight that shows the front of a vehicle. Unfortunately, the plate is extremely blurry and all normal sharpening tricks have failed. I need a forensic imaging specialist who can push well beyond routine filters—think de-blurring algorithms, AI-based super-resolution, frame averaging, or any other proven technique—to extract a legible plate number. Source material • One “very blurry” video frame pulled directly from the original recording. • Daytime lighting; no night-vision artifacts to worry about. What I would like back • A cleaned, high-confidence still image that clearly shows the license plate characters. • A brief note (a paragraph or two is fine) outlining the process and software you use...

Looking for a developer to build an iOS/Android app that helps couples and friend groups decide what to do locally. Core functionality: Collect local events, activities, restaurants, attractions, etc. from online sources/APIs Users create or join a couple/friend group Each person privately selects what they’re interested in that week App recommends a small shared batch of activities based on group interests Each person privately votes NO, LIKE, or LOVE When multiple people like the same activity, it becomes a match Show useful details such as location, date/time, price, directions, tickets/reservations, etc. Learn from past votes to improve future recommendations Push notifications for new recommendations and matches Basic admin tools to review/manage event data

Merhaba, Kitantik mağazam için basit bir ürün yükleme otomasyonu yaptırmak istiyorum. İstediğim sistem şu şekilde çalışacak: 1. Telefon veya bilgisayardan ürün/kaset fotoğrafını sisteme yükleyeceğim. 2. Fotoğraftaki renkli sticker algılanacak. 3. Sticker rengine göre daha önceden belirlediğim fiyat otomatik atanacak. Örneğin: - Kırmızı = 250 TL - Mavi = 200 TL - Yeşil = 150 TL - Sarı = 100 TL Bu fiyatları daha sonra panelden kendim değiştirebilmeliyim. 4. Yapay zekâ fotoğrafı analiz ederek mümkün olduğunca: - sanatçı - albüm/kaset adı - marka - ürün türü - fotoğrafta görülebilen diğer bilgiler bilgilerini çıkarsın. 5. Fotoğraf...

Busco Tech Co-founder para escalar dos plataformas IA en producción (Modelo Sweat Equity / Bootstrapping) Tengo dos plataformas ya en producción, traccionando y creciendo: * Un sistema multi-agente centrado en procesos de negocio. * Un motor de evaluación UE_IA basado en Inteligencia Artificial. Actualmente el proyecto se encuentra en fase de bootstrapping, por lo que busco un Tech Co-founder que se apunte a la aventura de escalar, mantener y evolucionar ambos productos conmigo. La prioridad inmediata es asegurar el rendimiento, la estabilidad y la capacidad de crecer en usuarios sin sacrificar la velocidad de entrega de nuevas funcionalidades. Tu día a día como socio tecnológico: Girará en torno a revisar el código existente, opti...

I want to put a complete, web-ready AI expert system online that suggests the best laptop for a user based on a few quick questions. The core of the project is a rule-based engine built with Experta in Python and wrapped in a Streamlit interface. Scope of work • Create or refine the rule base so that recommendations place the strongest weight on RAM and CPU, while still considering GPU, storage, OS and intended usage. • Design a simple, clean Streamlit front end—no clutter, just concise inputs and an easy-to-read results panel. • Load and preprocess a laptop data set (I can supply one, or you can extend it) and keep it in a Pandas-friendly format. • Return a ranked list of laptops with a brief rationale for each choice. Deliverables 1. Fully comment...

Looking to develop, integrate, and maintain an AI-operated proprietary fund under an LLC structure. The project specification: FSM logic sequencing of entry/exit state change signals for automation of TQR program trading and portfolio management algorithm. Ideal candidates must have expertise in program trading, portfolio management automation, FSM logic sequencing for entry/exit strategies, and state change signaling for automation. Dedication to this project is essential, and candidates should ensure it does not compete with other job commitments. The timeline for completion is 20 days.

we’re looking for engineers who know vllm and want to help us launch our product. we’re especially interested in people who are active on hacker news, reddit, github or product hunt and can bring interested developers and users to the launch. the more people you bring in who try the product, give useful feedback or become users, the bigger the reward. if this sounds like you, let’s talk.

I need an end-to-end algorithm that trades a single share-market instrument entirely on technical chart analysis. The core logic must follow a technical trading approach, relying on candlestick patterns, moving averages, RSI, and manually-defined trend lines to decide entries, exits, and position sizing. Scope The build starts from raw market data ingestion, moves through signal generation, and finishes with live order execution and performance reporting. All elements—data cleaning, feature engineering, back-testing, walk-forward validation, risk controls, and real-time monitoring—should sit inside one coherent codebase so I can flip from historical tests to live trading with minimal re-configuration. Implementation Notes • Python is my preferred language; common...

I need a complete solution that pinpoints abnormal behaviour across my wireless sensor network. You will have direct access to three rich data streams—raw sensor readings, detailed network-traffic captures, and accompanying environmental context—so the approach can fuse information rather than relying on a single source. Your job is to design, implement, and validate an anomaly-detection model that flags faulty nodes, suspicious traffic, or out-of-range measurements in near-real time. Feel free to select the technique that best fits the data volume and complexity; I am open to statistical, classic machine-learning, or deep-learning pipelines as long as the final system is transparent and reproducible. Python (NumPy, Pandas, Scikit-Learn, TensorFlow/PyTorch) is preferred for ...

I have a raw customer dataset and I need a predictive model that can automatically segment these customers into actionable groups. The job begins with cleaning and preparing the data, moves on to feature engineering, and culminates in training and validating a segmentation model in Python—ideally using pandas, scikit-learn (or similar), and clear visual output via matplotlib or seaborn. What matters most is accuracy and interpretability. I want to understand which features drive each segment, so well-commented code, concise documentation, and an executive-level summary of insights are essential. Final deliverables should include: • A reproducible Python script or notebook • The trained model (saved or easily retrainable) • A brief report highlighting key segments...

My aim is to develop a Machine Learning algorithm that can predict the top 4 finishers in horse, harness, and greyhound racing. Unfortunately, I do not have any historical data available for training the algorithm, so I will need a freelancer who has the expertise to build this from scratch. Ideal Skills and Experience: - Extensive knowledge and experience in Machine Learning. - Familiarity with racing dynamics, particularly horse, harness, and greyhound racing. - Ability to create synthetic data for algorithm training. - Prior experience in developing predictive algorithms. - Strong statistical analysis skills.

Senior AI/ML + Distributed Systems Engineer Required (Enterprise Build – $45,000 Fixed Price) Project: AI Financial Intelligence & Risk Monitoring Platform We are hiring a highly experienced engineer or engineering team to architect, build, deploy, and operationalize a production-grade AI Financial Intelligence & Risk Monitoring Platform. This is not a prototype. This is not a simulation. This must be a fully scalable, real-time, enterprise-grade system capable of operating under production conditions. ⸻ Project Overview The platform must: * Process high-volume financial data streams in real time * Detect anomalies using AI-driven models (not rule-based systems) * Generate actionable risk intelligence * Operate under strict performance, latency, and scalability benchm...

JOB: Computational Biologist / ML Engineer - Brain-Specific Neoantigen Prediction Tool PROJECT OVERVIEW: I am building a brain-specific neoantigen prediction tool for mRNA/LNP/PNA therapeutics. The goal is to predict which mutated peptides (neoantigens) are most likely to be presented by HLA in brain metastasis and trigger an immune response. This is a 7-stage pipeline with a defined step list. WHAT YOU WILL BUILD: Stage 1 - Data Sourcing - Search and download brain metastasis MS immunopeptidomics data from CPTAC, SysteMHC, GEO, PRIDE - Download TCGA primary tumour WES/RNA-seq (20 samples) - Download BrainMetShare brain metastasis WES/RNA-seq (20 samples) - Download GTEx normal brain expression, AFND HLA frequencies, IEDB self-antigens - Record sample metadata (source, cancer type, ance...

I’m building a computer-vision pipeline focused on reliable detection and recognition of human faces and full-body figures. The core of the job is a clean, well-documented Python implementation that can take still images or short video clips and return bounding boxes, class labels, and confidence scores for each detected person. I already have test media and the computing environment; what’s missing is the detection logic itself—ideally leveraging familiar libraries such as OpenCV, TensorFlow, PyTorch, or a proven YOLO/SSD variant. Accuracy on varied lighting and crowded scenes is more important to me than sheer speed, but the code should still run in real time on a modern GPU. Deliverables • Python source code with clear inline comments • Pre-trained weig...

I’m looking to turn our scattered know-how—hundreds of PDFs, policies, detailed reports, emails, meeting minutes, and assorted documentation—into a single Retrieval-Augmented Generation assistant that answers questions instantly and finds the exact source passage behind every reply. What I already have in mind • Ingestion pipeline that automatically pulls new or updated PDFs, email archives, and other internal files, cleans the text, and embeds it in a vector store (FAISS, Pinecone, or a similar tool). • A lightweight search layer so staff can type a query and receive both a concise answer and the ranked source documents. • Chat-style interface (web or Slack/Teams bot) backed by an LLM with citations, guardrails, and role-based access. • Modular...

My manuscript on temperature prediction with a Bidirectional LSTM has been returned from a journal with a “major revision” decision. I want to submit a polished, fully compliant revision that answers every reviewer comment and tightens both the science and the prose. The current file is in Word, so all editing should be done with Track Changes enabled. I will supply the reviewers’ reports, original figures, datasets, and any code snippets you may need to replicate or clarify results. Here is what I expect: • A carefully rewritten Word document that integrates the reviewers’ suggestions while preserving my voice and technical accuracy. • Clear explanations—inserted as comments—showing where and how each reviewer point was addressed. &bul...

Παρακαλούμε Κάνε Εγγραφή ή Κάνε Είσοδο για να δεις λεπτομέρειες.

I am looking for an experienced Python and Machine Learning developer to build an AI-based bias correction model for WRF 2-meter temperature (T2m) forecasts. Historical data will include WRF forecasts and observed temperatures from weather stations . The goal is to train a model to predict forecast bias and produce corrected temperature forecasts.

I run several AI initiatives that revolve around computer-vision pipelines, and I need an adaptable hand to keep our image data in perfect shape before it ever touches a model. Your main focus will be classic machine-learning groundwork: pulling raw image assets from cloud storage, auditing their quality, handling augmentations, labeling inconsistencies, and packaging everything into tidy, well-documented datasets that flow straight into our training scripts. Most tasks live in the data-preprocessing and cleaning stage, so you should be comfortable writing reproducible code for resizing, normalization, class-balancing, and automated sanity checks. We currently use Python with Pandas, NumPy, OpenCV, and sometimes Albumentations; if you have a favorite toolkit that speeds things up, I am op...

I’m looking for an experienced AI/Software Engineer to work on NL2Repo and repository-level coding tasks. The work involves translating natural-language requirements into changes to existing code repositories, including implementation, testing, debugging and validation. Before hiring, I need to review relevant previous work. Please send 1–3 strong samples showing experience with NL2Repo, SWE-bench, coding agents, repository-level code generation, automated code repair, or similar projects. Please include: * Your role in the project * GitHub/demo/patch if available * Technologies/tools used * Brief explanation of what you built Please don’t send unrelated portfolio work. I’m specifically looking for real repository-level engineering experience.

I am ready to commission a full-length, original research manuscript in artificial intelligence / machine learning that is suitable for submission to a Scopus-indexed, SCI-listed journal. At this stage I have no fixed problem statement, so I will rely on you to help shape a strong, publishable research question, design the methodology, run the experiments, and craft the paper end-to-end. Note : Apply only if you have a good publication record. Key deliverables • 2-3 complete manuscript (abstract through references) ready for journal submission • All code, datasets or preprocessing scripts required to reproduce the results • A concise response letter template explaining originality, significance and potential reviewer concerns Acceptance criteria will be clarity ...

I want to build an AI that can take any text prompt I feed it and return a lively, cartoon-style image. The core of the work is training or fine-tuning a model so it consistently produces bright, clean, character-driven artwork rather than photorealistic renders. I am open to the underlying architecture—whether that is a custom GAN, a diffusion approach, or a tailored Stable Diffusion pipeline—but the final results must read unmistakably as “cartoon” at first glance. To help me choose the right partner, send a detailed project proposal that walks me through your planned workflow: data collection or augmentation strategy for cartoon images, model selection, training schedule, expected hardware needs, and how you will package the finished system for everyday use (com...

I want to guide a small group of absolute beginners all the way to the point where they can land an entry-level role in AI, machine learning and generative AI. To get there, I’ll need a mentor who can design and deliver a complete learning journey—covering core theory just enough to ground them, then pivoting quickly into hands-on coding, model training and project work that showcases real-world skills. Here’s what I’m expecting from you: • A clear, modular curriculum that progresses from Python basics through classic machine-learning workflows and on to modern generative models (GPT-style LLMs, diffusion, etc.). • Live or recorded sessions with practical demos, Jupyter notebooks and datasets my trainees can keep experimenting with after class. •...

I have a clean, ready-to-use Iris dataset and need a complete Support Vector Machine classification pipeline built around it. The job is straightforward: tune an SVM (scikit-learn or similar) to separate the three Iris species, document the process, and hand over reproducible code and results. Here’s what I expect: • A concise notebook or script that loads the data, performs any minimal preprocessing you find beneficial, runs hyper-parameter optimisation, trains the final model, and outputs accuracy, precision, recall and the confusion matrix. • A short write-up (markdown inside the notebook is fine) explaining why the chosen kernel and parameters work best, plus any insights you notice in the feature space. • Saved model file so I can deploy or reload it quickl...

I need a standalone Windows application that bundles a local GPT-style language model with a voice layer powered by Google Assistant. The workflow is simple: I open any PDF, the model ingests it locally, and then—through voice alone—I can ask questions about the document and get spoken answers back. Key points I must see working: • Runs entirely on Windows without needing command-line setup. • Uses the Google Assistant voice interface for both input and spoken output. • Processes PDFs offline: the text is extracted, fed to the model, and retained only on my machine. • Voice queries such as “What is the author’s main argument on page 5?” return an accurate, concise spoken answer within a reasonable response time. I’m open to wh...

We are looking for freelancers from the USA and LATAM regions to participate in a large-scale egocentric (first-person point-of-view) video collection project focused on capturing genuine household and industrial activities. The collected footage will help train and improve AI/ML models related to human activity recognition, hand-object interaction, robotics applications, and real-world scenario understanding. Participants will be compensated at $10 per approved video hour for USA contributors and $7 per approved video hour for LATAM countries contributors. Freelancers will record themselves performing real-world household and industrial activities, including cleaning, organizing, maintenance, logistics, retail, agriculture, manufacturing, construction, automotive services, and other wo...

The project centers on building a production-ready TensorFlow 2.x model that classifies tabular data delivered to us through an internal API. I have the API specifications and sample payloads ready; you will turn those streams into a clean training pipeline, engineer the right features, and iterate until the classifier meets our performance targets in real-world tests. Scope of work • Data pipeline – pull the API data, handle preprocessing, and produce TensorFlow-friendly datasets for train/val/test splits. • Model development – design, train, and tune a deep learning architecture suitable for tabular inputs (e.g., wide & deep, Transformer, or other proven structures). • Optimization – experiment with hyperparameters, regularization, and callback...