Hire the Best PyTorch Freelancers
in Turkey

Top 3% Freelancer who built enterprise multi-tenant RAG systems with custom LLM deployments, reduced LLM workflow costs by 70%, and delivered a 2M product data quality pipeline for a Fortune 500 biotech company, reducing manual efforts by 85%. Architected and deployed production systems across finance, biotech, energy, telecommunications, and healthcare. 5+ years taking end-to-end ownership on all projects, from greenfield architecture to secure & scalable production deployment, leading engineers and maintaining systems long-term. I specialize in production-ready AI solutions with measurable ROI. From 90% accuracy voice classification to billion-scale data engineering. Core Expertise: • AI Agents / Agentic Workflows (LangGraph, Tool Calling, Function Calling, Agent Memory Patterns, Agent Patterns, Multi-Agent Orchestration, Building MCP, Plugins, Skills, Hooks, Deterministic Enforcement, ReAct Pattern, AI Assistants, Research, Planning, Writing) • LLM Systems (Claude API, OpenAI API Integration, vLLM, LoRA/QLoRA Fine-Tuning, PEFT, RLHF, DPO, LLaMA, Mistral, Prompt Engineering, Deterministic Workflows, LLM Evaluation, Cost Optimization, Deploying Chatbots, Deploying LLMs, Claude Code, Codex) • RAG (Hybrid Retrieval, Vector Databases, Evaluation Harness, Tiered Access, Custom Document Parsing, PDF/Clinical Data Extraction, Ragas, Multi-Tenancy) • NLP (Transformers, spaCy, Named Entity Recognition, Text Classification, Taxonomy Design, Rule-Based Extraction, Sentiment Analysis, Topic Modeling, Summarization, Information Extraction) • Generative AI (Structured Outputs, Pydantic Validation, Guardrails, Hallucination Prevention, AI Chatbot Development, AI Integration) • ETL & ML Pipelines • Time Series Applications at scale. - I personally prefer native APIs for building Agentic / LLM workflows. Domain Expertise: • Energy & Utilities (demand forecasting, carbon emissions, time series, pricing optimization) • Financial Services (compliance-ready RAG, churn prediction) • Life Sciences (biotech product classification, scientific NER, clinical data extraction, safety) • Telecommunications (voice classification, Whisper, ASR, diarization) Projects - Open source contributor (Langchain vLLM integration fix) - Enterprise RAG Engine with multi-tenancy, project-level isolation, custom vLLM deployments, agent live on the WhatsApp channel - customer support. - Finance-Specific RAG System with compliance features and RAG benchmarking - Enterprise Knowledge Base Chatbots Integrating (RAG daily sync on Notion, Slack, Website, Documents, evals, Live on Zendesk Customer Support Agent, Internal AI Agents on Slack(Deep Research, Q-A, Reporting), Blogpost creation with tone and fact checks.) - Production agentic workflows with state management, semantic processing, and monorepo CI/CD - SEO keyword clustering system processing 10k+ keywords under 10s using OpenAI embeddings + HDBSCAN - Enterprise Misspelling Detection Pipeline for 2M biotech products using FAISS similarity matching and 740k scientific corpus validation (Thermo Fisher) - Real-time Voice Classification System (90% accuracy, active learning pipeline) - Call Recommendation System with statistical engine for optimal calling times - Energy Market Forecast System for Budapest electricity prices - 500+ Class Text Classification with Fine-Tuned BERT - Extending a Lexicon using GloVe Word Embeddings - Databricks PySpark Data Pipeline for Aggregating Billion Scale Data - Time Series Analysis and Models on Energy Consumption Data from different locations in Belgium and Germany - API Integrations & Calculations on Electricity Business Objectives - Imbalanced Binary Classification on Anonymized Data (Oversampling + CatBoost & hGBM) - Data Processing Pipeline for 70m of Electricity Sensor Data for a Research Project Tech Stack: • Programming Languages: Python, SQL • Machine Learning & Deep Learning: TensorFlow, Keras, PyTorch, Scikit-Learn • NLP: LangChain, NLTK, Gensim, FAISS, Pinecone, pgvector • Time-Series: Prophet, TSMixer, ARIMA, SARIMAX, LSTM • MLOps & Data Engineering: Terraform, Linux, Apache Spark, Airflow, Docker, Flask, Streamlit, Gradio, MLflow, W&B, Optuna, CI/CD Pipelines, n8n workflows, Ragas • Cloud: Google (BigQuery, Cloud Functions, Cloud Run, Dataflow, Vertex AI, Firestore, Cloud Storage, Composer) • AWS (Lambda, SageMaker, S3, EC2, ECR, DynamoDB, Glue, CloudWatch) • Azure (Databricks-PySpark) • Database: PostgreSQL, Pinecone, Chroma • Data Manipulation & Statistics: Pandas, NumPy, OpenCV, Regex, SciPy, Statsmodels • Data Visualization: Matplotlib, Plotly • Web Scraping: BeautifulSoup, Requests • Building Custom NER Models & Linking, Question Answering, Text Similarity & Word Embeddings, Part-of-Speech Tagging & Dependency Parsing, Text Generation, Lexical Semantics & N-grams, Clustering • Logistic Regression, SVM, Decision Trees, Random Forest, Gradient Boosting Machines (XGBoost, LightGBM, hGBM), k-NN, CNN, RNN, K-Means, DBSCAN, HDBSCAN, OPTICS

WebRTC, FFmpeg, AI, OpenCV, Streamin – I fix real-time video apps and have since 2005. 13,485 hours, 14 projects, 100% Job Success. Zero abandoned. ⚡ 𝗙𝗿𝗲𝗲 𝗣𝗥𝗗. 𝗠𝗩𝗣 𝗶𝗻 𝟳-𝟭𝟰 𝗱𝗮𝘆𝘀. 𝗙𝗿𝗲𝗲 𝗰𝗼𝗱𝗲 𝗿𝗲𝘃𝗶𝗲𝘄 𝗼𝗿 𝟮-𝘄𝗲𝗲𝗸 𝘁𝗿𝗶𝗮𝗹 𝗼𝗻 𝗮𝗻 𝗲𝘅𝗶𝘀𝘁𝗶𝗻𝗴 𝗰𝗼𝗱𝗲𝗯𝗮𝘀𝗲. Netflix, HBO, EA, Nokia, the World Bank, and Harvard trust platforms I've built. These clients don't give second chances. Mine work. ⚠️ Invite me to your job – I reply within 15 minutes. 𝗦𝗲𝗹𝗲𝗰𝘁𝗲𝗱 𝗢𝘂𝘁𝗰𝗼𝗺𝗲𝘀 → Speed.Space – live production platform for Netflix, HBO, EA, and Apex Legends. Multi-stream switching, real-time collaboration, pro codec control. Used at Paris Fashion Week and Live Nation events. Zero downtime. → ShortKlips – video project management for Nokia and the World Bank. 200K+ monthly video views. Cut client project turnaround by 30%. → VirtualCr8tive – hybrid event platform for Harvard University and U.S. Vets. → CloudDoctors – HIPAA-compliant telehealth on WebRTC. Built from concept through multiple release cycles. → YUUALL – WebRTC platform built across multiple versions, long-term engagement. 𝗪𝗵𝗲𝗿𝗲 𝗜'𝗺 𝗦𝘁𝗿𝗼𝗻𝗴𝗲𝘀𝘁 – Pipeline latency – I profile FFmpeg filter graphs and cut processing time without losing quality. – WebRTC under load – jitter buffer, codec mismatch, TURN/STUN. I find which one is yours fast. – AI on live video – TensorFlow inference under 80ms; object and audio recognition on live feeds. – FFmpeg deep work – custom filter graphs, muxing, hardware acceleration (NVENC, VAAPI, VideoToolbox). – Scale & reliability – systems handling 600M+ call minutes a month, fixed at the architecture level. 𝗦𝗽𝗲𝗰-𝗙𝗶𝗿𝘀𝘁 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 I work spec-first, not vibe-first. I write the full spec and architecture before any code. Then I drive AI agents: developer, tester, analyst – against strict guardrails to produce production-grade code. The hard real-time work I handle or verify myself: WebRTC pipelines, sub-second latency, on-device ML. You get AI speed with the stability of a senior engineer, 4-10× faster than classic coding. 𝗛𝗼𝘄 𝗜 𝗪𝗼𝗿𝗸 You describe what's broken or what you're building. I send back a diagnosis: the problem, the fix, the cost. Not a sales call. Then we scope the work together before either of us commits. My estimates average ~10% off actuals. I validate before building and run staged QA on every release. 𝗛𝗼𝘄 𝗬𝗼𝘂 𝗖𝗮𝗻 𝗪𝗼𝗿𝗸 𝗪𝗶𝘁𝗵 𝗠𝗲 1. Build from scratch – real-time video, AI, or streaming, end-to-end. 2. Drop in on your team – I take the hard video, AI, or WebRTC modules. 3. Rescue and modernize – fix slow or unstable platforms and ship to production. I work inside the Fora Soft team – 51-200 engineers. If your project grows, a PM, QA, and more developers are already there. No solo freelancer who vanishes when scope expands. 𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝗶𝗲𝘀 𝗜 𝗨𝘀𝗲 𝗗𝗮𝗶𝗹𝘆 ◆ Real-time & video: WebRTC, LiveKit, FFmpeg, GStreamer, OpenCV, HLS/DASH, RTMP/RTSP, OBS, WebGL ◆ AI & ML: TensorFlow, PyTorch, OpenAI API, Whisper, YOLO, object and audio recognition ◆ Back-end: Node.js, Python, TypeScript, distributed systems, high-load APIs ◆ Cloud & infra: AWS, GCP, Docker, Kubernetes, CI/CD 𝗪𝗵𝘆 𝗙𝗼𝗿𝗮 𝗦𝗼𝗳𝘁 The team behind me: 914 jobs · $10M+ earned · 401K+ hours · 250+ projects since 2005 · 400+ clients across 17+ countries · Top WebRTC & Telecom Developer 2022 and 2024 · 1 of 400 devs passes our internal selection. 𝗖𝗹𝗶𝗲𝗻𝘁 𝗙𝗲𝗲𝗱𝗯𝗮𝗰𝗸 "We enjoyed working with Andrey. He saw himself as part of our team and proactively contributed. Even last-minute requests were handled in a timely, constructive way. Many of his proposals were better than our own ideas." – client, YUUALL (WebRTC platform) ⚠️ Describe what's broken. I'll tell you what's wrong, how to fix it, and what it costs.

I engineer end-to-end spatial data pipelines and custom WebGIS platforms. My focus is on high-level geospatial architecture—solving structural bottlenecks to build repeatable workflows, clean spatial databases, and practical mapping systems that operational teams can actually use. Recent US project work includes address matching and territory planning for operational teams, public facing GIS maps for policy and exhibition use in Omaha, property and building height analysis in Los Angeles, and a browser based weather graphics platform built with MapLibre, radar layers, timeline controls, and export ready map views. International work includes site selection analysis in Germany, parcel and location intelligence workflows, PostGIS and QGIS topology cleanup, UK air quality grid matching, raster processing, and WebGIS tools built with MapLibre, Google Maps API, AWS, and spatial databases. I can support projects involving GIS analysis, WebGIS dashboards, PostGIS workflows, raster and remote sensing analysis, automated geospatial pipelines, and custom mapping tools. If you need a one time analysis, a repeatable GIS workflow, or a full WebGIS prototype, I can take the project from raw data to final deliverable with a clear technical process.

I'm a computer engineer fresh graduated also i am industrial engineer (major degree) and i want to improve myself while i am working as a freelancer websites. You can reach me via linkedin: in/ernsmsr/ /github/cynorra/

MSc. Data Scientist with 3+ years of experience in transforming raw data into valuable services. Proven ability to train ML models(classification, clustering, regression) and implement them into pipelines for inference in real-world applications by integrating them into Restful APIs, and finishing the process by Dockerizing the service to be used in production. 💻𝙒𝙃𝙀𝙍𝙀 𝙄 𝙀𝙓𝙋𝙀𝙍𝙏𝙄𝙎𝙀: - Data science: Data visualization, Supervised learning algos, Unsupervised learning algos, Regression, ANN, Feature Engineering, Feature selection & extraction, etc.. - Computer vision: Object Detection(YOLO, Faster-RCNN), CNN architectures(Resnet, EfficientNet), Zero-shot learning(CLIP, GDINO, SAM, BLİP), VIT - LLM: LangChain, LangGraph, RAG systems, model quantization, LORA, LLM architectures(BERT, LLAMA), OpenAI API, HuggingFace Transormers - ML libraries: Pandas, Numpy, Keras, Pytorch, Tensorflow, AutoML, MLflow - Programming Languages: Python, Dart, SQL, NoSQL - Databases: Vector DB(Chroma, PGvector, LlamaIndex), PostgreSQL, Elasticsearch - Backend: Fast API, Flask - Cloud Platforms: AWS(Sagemaker, S3), GCP(Vertex, vision API), Kubernets, Docker, Github - Operating Systems: Linux, Windows 💡𝙒𝙃𝘼𝙏 𝙔𝙊𝙐 𝙒𝙄𝙇𝙇 𝘽𝙀 𝙋𝙍𝙊𝙑𝙄𝘿𝙀𝘿: Build advanced AI models and seamlessly integrate APIs to enhance workflows. Collect, process, and analyze data to reveal valuable business insights. Leverage machine learning and data-driven solutions to automate and optimize tasks. Offer dedicated long-term support for sustainable AI and data-driven initiatives. 🔎𝘾𝙊𝙉𝙉𝙀𝘾𝙏 𝙒𝙄𝙏𝙃 𝙈𝙀, 𝙄𝙁: You're seeking a seasoned expert in AI and API integration to drive business innovation. You need efficient data scraping and analytics solutions to inform critical decisions and strategy. You aim to harness cutting-edge technology to boost productivity and maximize profits. You require a dedicated specialist to manage, streamline, and enhance your tech infrastructure for optimal performance. 🚀𝙏𝙃𝙀 𝘼𝘿𝙑𝘼𝙉𝙏𝙀𝙂𝙀𝙎 𝙊𝙁 𝘾𝙊𝙇𝙇𝘼𝘽𝙊𝙍𝘼𝙏𝙄𝙉𝙂 𝙒𝙄𝙏𝙃 𝙈𝙀 : Fast and professional service centered on delivering client success. Tailored insights and cutting-edge solutions for your data and AI challenges. Committed long-term support to drive ongoing improvement and business growth. Reach out on Upwork to explore how I can help you harness data and AI for your business success!

A proficient freelancer in MATLAB, Python, React Native, C, C++, Keras, PyTorch, TensorFlow, image processing, data classification, AI, ML, DL and Scrapping. My expertise spans a wide range of computer skills, including: - Languages/Technologies: MATLAB, Python, React Native, Java, C, C++ Etc. - Machine Learning / Deep Learning: Keras, Pytorch, Tensorflow - Simulation: Pspice, Proteus, Zemax Designer - O.S & Tools: LaTeX, Microsoft Office, Linux (Ubuntu & Manjaro) - Databases: MySQL, MsSQL - Network Technologies: TCP/IP, HTTP/HTTPS, OSI Model - Web Services: Apache, NGINX - IoT Communication Protocols: MQTT, LoRaWAN, Bluetooth, ZigBee, WiFi With expertise in AI and ML, I am at the forefront of driving advancements in the following areas: - Wireless Communications: MIMO, NOMA, mmWave, beamforming, user pairing, multicarrier technologies, spectrum management in IoT, energy efficiency, and wireless sensor networks. - Signal Processing: Image processing, biomedical signals, passive sonar data, and digital signals. - Holographic Beamforming for Massive MIMO. - Interference Management in cooperative network systems. - Visible Light Communication in MIMO systems. Translation, Transcription, Proofreading Services: English <=>Persian English <=>Dari English <=>Turkish English <=>Pashto