Thrilled to have conducted three back-to-back Learning & Development workshops at Christ University, Bangalore on Thursday, with over 600+ enthusiastic participants joining us! It was a fantastic opportunity to share insights, foster growth, and connect with bright minds eager to learn and evolve. ❤️ A special congratulations to the standout performers from the sessions: Sahil Yadav Anjali Odayoth Bhavika Sundar Evangeline Praisy Amana Susan Saji kingchum zongbu lepcha Shreya R Your enthusiasm and commitment truly set you apart. A huge thank you to the Christ University for inviting the Soul AI team, and for their warm hospitality—and, of course, the thoughtful token of appreciation! 🙏 Here’s to the next generation of AI enthusiasts and community builders. 🚀 #ArtificialIntellegence #ProfessionalDevelopment #Learninganddevelopment #ChristUniversity #SoulAI #LearningJourney
About us
We are an Enterprise AI firm focusing on RLHF and Custom AI solutions. We are dual incorporated in the US and India (our India office is in Gachibowli, Hyderabad). We are led by an IIT-IIM founding team, and have a highly skilled talent pool. Think PhDs, Engineers, Artists. We value wellness. Happy Employees = High Quality Work :) Our startup DNA helps us execute at the speed of light. Apply for our remote AI trainer opportunities at https://app.soulhq.ai
- Website
-
https://www.soulhq.ai/
External link for Soul AI
- Industry
- Software Development
- Company size
- 11-50 employees
- Headquarters
- Delaware
- Type
- Privately Held
- Founded
- 2023
- Specialties
- Artificial Intelligence, Generative AI, LLMs, Deep Learning, RLHF, SFT, and Data Annotation
Locations
-
Primary
Get directions
Delaware, US
-
Get directions
Financial District, Gachibowli
Hyderabad, IN
Employees at Soul AI
-
Dhanikh ..
Growth | B2B SaaS Product Management | Marketing
-
Gaurav Kumar
Contributing in the field of Structures Applications in Aerospace Sector. Engineer at heart, Solver by profession and Entrepreneur in mind. Lets talk!
-
Dr Nazia Khan (PhD)
Generative AI Expert/ Trainer and Evaluator (STEM) | Co-founder and Joint Secretary @ Anwar Hockey Society
-
CA Nagarjun G.
Partner @ Raj Niranjan Associates| Auditing, Bookkeeping, GST Consultant| Freelancer @ Upwork, Soul AI|
Updates
-
Soul AI is thrilled to announce that we will be visiting Christ University on 10th October to conduct an exclusive workshop: "Mastering AI Skills 101". 🚀 This is a great opportunity for all AI enthusiasts, tech aficionados, and students eager to dive deeper into the world of Artificial Intelligence. Whether you're just starting or looking to refine your AI skills, this workshop is designed to provide valuable insights and hands-on learning. 👉 When: 10th October 👉 Where: Christ University Campus 👉 What: AI Skill Mastery, Networking, and more! If you're passionate about AI or simply curious to learn more, we’d love to see you there. Let’s connect, learn, and grow together! 🌟 Christ University, Bangalore Saksham Choudhary Vedank Bhatnagar Muskan Gupta #artificialintelligence #innovation #technology #christuniversity #SoulAI #AI
-
-
At Soul AI, we're passionate about making AI knowledge accessible to everyone. As the industry continues to grow, understanding the nuances of AI and effective prompt writing is crucial. Our workshops are designed to provide a comprehensive introduction to the AI landscape, focusing on how to leverage prompting techniques effectively. The AI & Prompting 101 workshop is a two-day program focused on deepening participants' understanding of Supervised Fine-Tuning (SFT) and Reinforcement Learning with Human Feedback (RLHF). It covers the essentials of Generative AI, Model Training, and advanced Prompting techniques, providing a comprehensive view necessary for working with Large Language Models (LLMs). Participants learn to develop complex prompts and create high-quality RLHF and SFT datasets, essential for fine-tuning LLMs. The workshop includes extensive hands-on experience, enabling attendees to apply advanced techniques in AI model manipulation, with a focus on precision in prompt engineering and detailed model fine-tuning. #prompting #artificialintelligence #datascience #AI #innovation #workshop
-
"Prompt engineering is the art of communicating with a generative large language model." Here at Soul AI we believe in learning while working, and upskilling ourselves constantly. 🚀 This is a very concise list of Prompt Engineering principles that will guide you to be more precise with your prompts. Reference: Paper written by Sondos Mahmoud Bsharat, Zhiqiang Shen and Aidar Myrzakhan Article: https://lnkd.in/e4KxQ5F6
🖊️ Prompt Engineering Principles Good prompting can be a complete game-changer in your applications. Here's a list of 26 principles to craft high-quality prompts, based on experiments with Llama 1 & 2 and GPT-3.5 & 4. Nothing new, but it's a good reminder that helped me today so I wanted to share it. It comes from a paper written by Sondos Mahmoud Bsharat, Aidar Myrzakhan, and Zhiqiang Shen. 📝 Article: https://lnkd.in/e4KxQ5F6
-
-
Soul AI reposted this
🚨🚀 We’re excited to unveil our contribution to open-source SLMs for the insurance sector. ✨ Presenting "phi-2-insurance_qa-sft-lora"! A 2.7B parameter, first of its kind SLM, capable of answering various queries around insurance in the American context. At Soul AI, we’ve helped ship Foundation LLMs for enterprises using RLHF. This release showcases our broader capabilities and reinforces our firm belief in pushing the frontiers of technical LLMs with the help of high quality expert data. 🦾 As the name suggests, the model is based on Microsoft's Phi-2, fine-tuned using the LoRA paradigm. 📊 Focused on Q&A in insurance, it taps into the depth of the InsuranceQA dataset, comprising 27.96K expert-curated QA pairs. The choice of the dataset enables the model to understand real-world queries, and reduces hallucinations. 🤖 We’re committed to open-sourced AI and are releasing the model under the MIT license. It is free for exploration, modification, and integration. We humbly invite the AI community to engage and collaborate with us on our open-source journey! 🔜 Next up? We’ll be bringing more interesting research, insights, models and datasets to your timelines! Assets: (A) Find the model collection at https://lnkd.in/g2r2V9hd. We provide merged weights in (1) safetensors, and (2) gguf formats. (B) Find the dataset in a readable `jsonl` format at https://lnkd.in/g7asWnRT. Citations: [1] Hu, Edward J., et al. "Lora: Low-rank adaptation of large language models." arXiv preprint arXiv:2106.09685 (2021). [2] Feng, Minwei, et al. "Applying deep learning to answer selection: A study and an open task." 2015 IEEE workshop on automatic speech recognition and understanding (ASRU). IEEE, 2015. [3] Mojan Javaheripi, Sébastien Bubeck. "Phi-2: The surprising power of small language models" (https://lnkd.in/gW7B-xMY) #LLM #Phi2 #SFT #RLHF #GenAI #OpenSourceAI #OSS #LoRA #FineTune #NLP #AI #SLM
-
🚨🚀 We’re excited to unveil our contribution to open-source SLMs for the insurance sector. ✨ Presenting "phi-2-insurance_qa-sft-lora"! A 2.7B parameter, first of its kind SLM, capable of answering various queries around insurance in the American context. At Soul AI, we’ve helped ship Foundation LLMs for enterprises using RLHF. This release showcases our broader capabilities and reinforces our firm belief in pushing the frontiers of technical LLMs with the help of high quality expert data. 🦾 As the name suggests, the model is based on Microsoft's Phi-2, fine-tuned using the LoRA paradigm. 📊 Focused on Q&A in insurance, it taps into the depth of the InsuranceQA dataset, comprising 27.96K expert-curated QA pairs. The choice of the dataset enables the model to understand real-world queries, and reduces hallucinations. 🤖 We’re committed to open-sourced AI and are releasing the model under the MIT license. It is free for exploration, modification, and integration. We humbly invite the AI community to engage and collaborate with us on our open-source journey! 🔜 Next up? We’ll be bringing more interesting research, insights, models and datasets to your timelines! Assets: (A) Find the model collection at https://lnkd.in/g2r2V9hd. We provide merged weights in (1) safetensors, and (2) gguf formats. (B) Find the dataset in a readable `jsonl` format at https://lnkd.in/g7asWnRT. Citations: [1] Hu, Edward J., et al. "Lora: Low-rank adaptation of large language models." arXiv preprint arXiv:2106.09685 (2021). [2] Feng, Minwei, et al. "Applying deep learning to answer selection: A study and an open task." 2015 IEEE workshop on automatic speech recognition and understanding (ASRU). IEEE, 2015. [3] Mojan Javaheripi, Sébastien Bubeck. "Phi-2: The surprising power of small language models" (https://lnkd.in/gW7B-xMY) #LLM #Phi2 #SFT #RLHF #GenAI #OpenSourceAI #OSS #LoRA #FineTune #NLP #AI #SLM
phi-2-insurance_qa-sft-lora - a soulhq-ai Collection
huggingface.co
-
We recently visited IIT Hyderabad for a student connect. In a very engaging session, we discussed how GenAI will shape the careers of GenZ. As AI becomes increasingly talented, the workforce of the coming generations will have to learn to operate in a world filled with promise but also with VUCA (Volatility | Uncertainty | Complexity | Ambiguity). As a part of our mission, we intend to reach every Indian student and help them improve their domain skills while training them to be better at prompt engineering. Onwards! 🚀 Special mention to Malaaika Chhaya and Gaurav kumar Singh for helping with the smooth execution of this session. Office of Career Services, IIT Hyderabad #genai #rlhf #genz #campusvisit #vuca #promptengineering
-
-
Excited to share that our revamped website is now up! Check us out at https://www.soulhq.ai/ 🚀
Soul AI | The Power of Humans in the Age of AI
soulhq.ai
-
Thank you for sharing this Jewel Syriac! 🙌 We love this analogy, as it absolutely encapsulates what RLHF is all about in an ELI5 manner!
AI trainer & experimentalist | Physics Expert | Online educator for worldwide students | Former Assistant Professor | Self taught Web Developer |
When I worked as an AI trainer 💻 at Soul AI, one of the challenging topics ⚠ I encountered was Reinforcement Learning from Human Feedback (RLHF). I came across an analogy 💯 that helped me understand it more clearly, and I want to share that with you 👨💻. Imagine teaching a child 👶 to speak 🗣 and think 🗯. You'd start by introducing them to basic language 🈹 🈲 through stories 🌃 and conversations 🗺, correcting them gently 🤝, and encouraging their curiosity 👌 . As they grow 👧 , you'd offer more complex ideas 🕵♀️ and nuanced feedback 🎯 , guiding them to understand not just words, but also the values ⛳ and context ♥ behind them. This is quite similar 👯♀️ to Reinforcement Learning from Human Feedback (RLHF) 🤖 in the AI world 👨💻 . Mainly RLHF contains: ▶ Supervised Fine-Tuning (SFT): Think of this as the initial schooling phase 🏫 for an AI. Just as a child learns basic language skills from books 📖 and conversations 👥 , the AI is fed a vast array of text 🔠 . This builds its foundational 🥅 understanding of language 🔤 . ▶ Reward Modeling: Here, AI trainers act like attentive parents 👨⚖️ or teachers 👩💼 , evaluating the AI's 'responses' or 'answers'. They rate ✔ these responses to teach the AI what a good answer looks like. It's similar to a parent praising 👌 a child for a thoughtful answer or gently correcting ❌ them when they're off-mark 🏸. ▶ Proximal Policy Optimization (PPO): This is where the AI starts 'practicing' 🏋♂️ on its own, similar to a child doing homework ⛹♀️ or solving puzzles 🧠 . The AI tries out different responses 🎰 , learning from the reward model which ones are good (get high scores 🎯 ) and which ones aren't 🚫 . It's a bit like a child learning from their successes and mistakes, constantly improving 🧗♀️ . ▶ Human-in-the-loop (HITL): The learning process for AI 👾 , like for a child 🚼 , is never done 🔁 . Human trainers stay involved, continually guiding the AI. This ensures that the AI's responses are not only correct ✅ but also appropriate ☑ and sensitive ⁉ to complex human standards. It's like a parent or teacher providing ongoing guidance and feedback as a child grows into an adult 🕺 💃 . So, RLHF in AI 🤖 is very much like raising or educating a child 🤓 . It starts with basic learning, moves to guided practice and feedback, involves learning from both success and failure and requires ongoing mentorship. This method helps AI or LLMs not just to 'know' things 🖥, but to understand and align with human values and context, much like how we educate children to become thoughtful, informed adults 👩🎓.
-
-
The Indian Republic Day might be over, but we aren’t over India’s thrilling role in today’s age of AI. So, we want to take a moment and talk about some of our favourite #IndianAI initiatives. Shoutout to Krutrim, bursting right off the blocks and becoming a #Unicorn earlier this week! Special mention to AI4Bhārat for constantly pushing the #SoTA in the Indic #NLP ecosystem. Over the last few years, they've released best in class Indic Neural Machine Translation / Natural Language Understanding models to the realm of #OpenSourceAI. Giving everyone a run for their money is Tech Mahindra's Project Indus! Their vision to create LLMs for Indian Languages has got the Indian #GenAI space hot. Give them a follow to witness the future! We would also like to humble brag about what we’re doing here at Soul AI and our own humble contribution to India’s role in AI. Be it Expert #RLHF or Supervised Fine Tuning or Enterprise AI, we are operating at the bleeding edge of things. Excited for what more is coming in 2024! #IndianRepublicDay #SoulAI #AI #RLHF #LLMs #GenerativeAI #NLP #ML