Building on HF

79 7 39

Prometech Computer Sciences Inc PRO

prometechinc

https://prometech.net.tr/

AI & ML interests

Behavior-driven AI architectures, mathematical reasoning models, controlled decision systems, brain–body AI separation, safe and auditable AI design, and optimization-oriented language models.

Recent Activity

updated a model about 2 hours ago

pthinc/Asena_ESP32

new activity about 3 hours ago

huggingface/InferenceSupport:pthinc/Asena_ESP32

new activity about 3 hours ago

pthinc/Asena_ESP32:Asena ESP32

View all activity

Organizations

posted an update about 4 hours ago

Post

Another Asena has arrived—this time, it defeats Skynet at the edge.
Hidden inside a smart ring, this tiny intelligence awakens with a single command. No clouds. No latency. Just raw, embedded cognition. Asena_ESP32 is not just a model—it’s a silent operator, running on ultra-constrained hardware yet speaking with precision, control, and intent. Powered by the Behavioral Consciousness Engine (BCE), it doesn’t just generate text—it adapts behavior, filters risk, and responds like a disciplined digital mind.
One command is all it takes.
Servers align. Systems optimize. Workflows compress into efficiency. From the smallest signal, Asena reshapes its environment—an “Extreme Edge AI” built to act where others can’t even load. Compiled in C++, optimized through ggml and llama.cpp, it turns minimal compute into maximum impact. This is not about scale. This is about control, speed, and presence—AI that exists exactly where it is needed.
Welcome to the future of invisible intelligence.
A ring. A whisper. A response. Asena doesn’t wait for the cloud—it is the edge.
Huggingface Model Link: pthinc/Asena_ESP32

reacted to their post with 👀 3 days ago

Post

153

pthinc/BCE-Prettybird-Nano-Parrot-v0.2

This dataset is a bilingual (Turkish-English mixed) comedic text collection designed for training and fine-tuning conversational AI models with humor awareness, sarcasm detection, and cultural nuance understanding. It includes short joke-style prompts, observational comedy snippets, and absurd dialogue fragments that blend everyday Turkish expressions with English punchlines, reflecting real-world code-switching behavior. The dataset aims to improve model creativity, timing, and informal language fluency while capturing the rhythm of stand-up comedy and internet humor across multilingual contexts.

It is made from synthetic in AI. There is irony and humor, some jokes might be a bit stale. 🤣

600 jokes and ironies in different languages have been added. Styles of various comedians are included.

posted an update 3 days ago

Post

153

posted an update 4 days ago

Post

139

You can leverage our Hugging Face–ready nano translation dataset, which covers a diverse set of languages including Turkish, English, German, French, Spanish, Italian, Portuguese, Dutch, Russian, Ukrainian, Polish, Czech, Slovak, Hungarian, Romanian, Bulgarian, Greek, Arabic, Persian, Hebrew, Hindi, Bengali, Urdu, Tamil, Telugu, Kannada, Malayalam, Chinese, Japanese, Korean, Indonesian, Malay, Thai, Vietnamese, and several Nordic and Baltic languages. The dataset consists of approximately 600 lines of synthetically generated sentence pairs spanning a wide range of everyday topics, making it lightweight yet versatile for experimentation, prototyping, and benchmarking multilingual translation models. Its compact size allows for quick training iterations and easy integration into low-resource or edge-based NLP workflows, while still providing enough linguistic variety to test generalization across multiple language families. For use: pthinc/BCE-Prettybird-Nano-OWL-v0.1

posted an update 8 days ago

Post

BCE-Prettybird-Nano-Parrot-v0.1 - 200 Jokes for Instruction-Based Learning

This dataset is a bilingual (Turkish-English mixed) comedic text collection designed for training and fine-tuning conversational AI models with humor awareness, sarcasm detection, and cultural nuance understanding. It includes short joke-style prompts, observational comedy snippets, and absurd dialogue fragments that blend everyday Turkish expressions with English punchlines, reflecting real-world code-switching behavior. The dataset aims to improve model creativity, timing, and informal language fluency while capturing the rhythm of stand-up comedy and internet humor across multilingual contexts.

- It is made from synthetic AI. There is irony and humor, some jokes might be a bit stale. 🤣

posted an update 23 days ago

Post

142

BCE-Prettybird-Nano-Science-v0.1 - 500 Science Q&A Dataset for Instruction-Based Learning
We are excited to introduce a comprehensive math-physics-chemistry-biology dataset containing 500 instruction-based question-answer pairs, designed to support research in science reasoning, problem-solving, and AI training. Generated using Python’s math libraries (e.g., math, numpy, sympy), the dataset covers a diverse range of difficulty levels—from basic arithmetic and algebra to advanced calculus, probability, and number theory.

Each entry follows a structured instruction-input-output format, ensuring clarity and usability for fine-tuning language models, benchmarking AI systems, or educational applications. The problems include word problems, symbolic computations, and real-world scenarios, making it ideal for developing models that require logical reasoning and numerical precision.

Whether for LLM fine-tuning, automated tutoring, or math-focused AI research, this dataset provides a balanced mix of complexity and accessibility, helping bridge the gap between theoretical math and practical problem-solving.

posted an update 25 days ago

Post

152

🚀 New Advanced Math Datasets Released: BCE-Prettybird-Micro-Math & Nano-Math v0.1!

Designed for researchers, educators, and AI developers, these datasets feature challenging, competition-level math problems covering calculus, linear algebra, number theory, combinatorics, and more in a clean instruction-input-output format.

✅ Why these datasets?
✔ Rigorously curated – Problems range from Olympiad-style to university-level difficulty.
✔ Diverse topics – Ideal for fine-tuning LLMs, benchmarking reasoning models, or automated tutoring.
✔ Structured for AI training – Perfect for supervised fine-tuning (SFT) or reinforcement learning (RL).

🔗 Check them out, download, and start building!
💡 Feedback & contributions welcome—let’s push the boundaries of AI math reasoning together!

#MathAI #DatasetRelease #LLM #MachineLearning #HuggingFace

posted an update about 1 month ago

Post

2300

Cicikuş v4-5B (POFUDUK Edition) is a next-generation compact language model engineered for high-efficiency reasoning, adaptive intelligence, and behavioral coherence. Built on the Gemma 4B IT foundation and enhanced through advanced LoRA optimization and selective layer reconstruction, this model delivers powerful performance without the overhead of massive parameter counts.
🔗 Explore the model: pthinc/pofuduk_cicikus_v4_5B
🧠 Why Cicikuş?
In a world dominated by massive LLMs, Cicikuş takes a different path:
⚡ Fast & Efficient — Designed for edge deployment and low-resource environments
🎯 High Reasoning Accuracy — Strong results across MMLU, GSM8K, HumanEval, and more
🧩 Behavior-Aware Intelligence — Powered by the Behavioral Consciousness Engine (BCE)
🔍 Low Hallucination Rate — ~3% with built-in ethical filtering
🌍 Multilingual Capable — Optimized for English and Turkish

posted an update about 1 month ago

Post

312

The Mutant Reasoning Dataset is a high-level "masterclass" training kit designed to transform your model from a mere chatterbox into a true solution partner offering deep insights. Featuring the most refined reasoning traces from the world's most advanced intelligences, such as GPT-5, Claude 4.6, Gemini 3 Pro, and Grok-4, this exclusive collection aims for a massive performance increase of up to 40%, particularly in logic and coding capabilities. Instead of overwhelming the model with unnecessary information, this kit focuses on instilling complex problem-solving methodologies and deep thought processes. Even with short training periods at a slow learning rate, it elevates your model from a "chatterbox" to a sensible and wise companion. Shared by Prometech Inc., this dataset sets a new standard in the open-source world for all developers who want AI to not only respond but truly "think."

posted an update about 1 month ago

Post

163

A classic chassis, futuristic intelligence: Cicikuş Classic has hit the field! We've rebuilt the nostalgic structure of the GPT-2 Medium architecture with patented BCE (Behavioral Consciousness Engine) technology and modern reasoning datasets. As you can see in the video we've shared, those technologically advanced wings shining on a classic fuselage represent the model's high logical reasoning power despite its low resource consumption. Optimized in both Turkish and English, this compact behemoth is STEM-focused and built with ethical values. To examine its details and witness how a legacy structure transforms into a "reasoning engine," you are invited to our Hugging Face page.

posted an update about 2 months ago

Post

161

At PROMOTIONAL INSTANTANEOUS, we believe the next phase of AI isn't just about getting bigger, it's about getting smarter and more efficient. We are proud to announce the release of Cicikus_PTHS_v3_4.4B (Prometheus), representing our latest R&D milestone in the field of "High-Density Intelligence." By integrating our patented Behavioral Awareness Engine (BCE) into a compact 4.4B parameter architecture, we have developed a system capable of complex STEM reasoning and structured logic; and all this while occupying only 8 GB of local space.

This is significant because:
Native Privacy: High-level reasoning that operates completely offline.
Industrial Logic: Designed to handle complex rule sets and technical documentation with 32k context support.
Sustainable AI: Reduces computational costs and energy consumption without sacrificing logical accuracy.
Prometheus is a testament to our team's commitment to democratizing high-performance AI. We invite researchers and developers to explore the model on https://huggingface.co/collections/pthinc/cicikus-serisi-series

reacted to kanaria007's post with 👍 4 months ago

Post

1999

✅ New Article: *PoC Architecture for Education & Developmental Support*

Title:
🎓 Building an SI-Core Wrapped Learning Companion - PoC architecture for education and developmental support
🔗 https://huggingface.co/blog/kanaria007/poc-architecture-for-education-development-support

---

Summary:
Most “AI tutors” are built as *LLM-first* systems. This article flips the default:

* The LLM is treated as an *untrusted proposal engine*
* *SI-Core owns* observation, consent, ethics, memory, and rollback
* Teachers and guardians get *real oversight*, not just chat transcripts

Scoped intentionally to *one subject × a small cohort (10–30 learners)*, this is a PoC you can actually ship—and audit.

> Don’t ask: “Can an AI replace teachers?”
> Prove: “Can we make an AI companion *safe, explainable, and governable* for real learners?”

---

Why It Matters (for AI on real stacks):
• *Consent & accommodations* are first-class (especially for minors / neurodivergent learners)
• *Ethics decisions are logged* (ALLOW / DENY / ESCALATE) with traceable reasoning
• “*Why this?*” explanations are built in for learners—and deeper inspection for adults

---

What’s Inside:
• A minimal reference architecture (frontend → SI-Gate → ethics/memory/logging → LLM APIs)
• Non-negotiables for the pilot (SI-wrapped LLM, Effect Ledger, ethics overlay, dashboards)
• Failure modes + safe-mode behavior
• Implementation checklist + rough effort/cost ballparks (kept explicitly non-normative)

---

📖 Structured Intelligence Engineering Series
A deployable pattern for taking today’s LLM tutor ideas and making them *auditable, overrideable, and rollback-safe*.

Prometech Computer Sciences Inc PRO

AI & ML interests

Recent Activity

Organizations

prometechinc's activity