Ähdus Technology logo
Ähdus TechnologyPosted 1 month ago

AI/ML Engineer Speech, RAG & Fine-Tuning

On-siteRawalpindi, Punjab, Pakistan

Full TimeSmall

Job Summary

Design and implement speech-to-speech pipelines using open-source models like Whisper and Wav2Vec. Develop and optimize STT and TTS systems with Coqui frameworks while applying LoRA and PEFT fine-tuning to LLaMA 2 and LLaMA 3 for domain-specific NLP tasks. Build RAG-based systems for knowledge-grounded responses and integrate agentic AI systems with reasoning capabilities. Collaborate with cross-functional teams to deliver scalable AI solutions and monitor deployed models for accuracy and efficiency. Full-time onsite role (10AM - 7PM) in Bahria Town, Rawalpindi.

Required Qualifications

  • Strong experience in AI/ML model development with open-source speech and language models
  • Hands-on experience with Whisper, Wav2Vec, Coqui TTS/STT frameworks
  • Proven track record with LLaMA 2, LLaMA 3 or similar LLMs
  • Proficiency in fine-tuning techniques: LoRA, PEFT, and parameter-efficient training
  • Experience in RAG-based systems for knowledge retrieval and contextual response generation
  • Familiarity with agentic AI frameworks for building task-oriented agents
  • Strong programming skills in Python, PyTorch, TensorFlow
  • Experience with Hugging Face, LangChain, and vector databases (FAISS, Pinecone, Weaviate, etc.)
  • Knowledge of cloud platforms (AWS, GCP, Azure) and containerization (Docker, Kubernetes)
  • Strong problem-solving skills and ability to optimize model performance

Desired Qualifications

  • Masters or PhD in Computer Science, AI/ML, Data Science, or related field
  • Publications or projects in speech AI, LLM fine-tuning, or agentic AI
  • Experience with distributed training and model deployment at scale

Hiring someone like this?

Get your role in front of qualified candidates on Sorce.

Get started

Apply to this job in one click with Sorce

Apply on Sorce