Switch Communications is Hiring AI/ML Engineer (LLM + Speech Systems)
Location: Onsite | Islamabad
๐๐๐ฒ ๐๐๐ฌ๐ฉ๐จ๐ง๐ฌ๐ข๐๐ข๐ฅ๐ข๐ญ๐ข๐๐ฌ:
1) AI Agent Development
* Build and deploy intelligent AI voice agents using open-source and custom models.* Develop STT, NLU, Dialogue Management, LLM, and TTS pipelines.
Build AI agents using frameworks such as Strands Agents, LangChain, LangGraph, or similar
Fine-tune LLMs such as LLaMA, Mistral, Falcon, etc.
Optimize agents for real-time, low-latency production use.
2) RAG & Model Training
Build and optimize RAG pipelines using vector databases, embeddings, chunking, retrieval, and context management
Prepare and manage English, Urdu, and regional language datasets.
Fine-tune models such as Whisper and Wav2Vec2 for speech recognition.
* Implement TTS solutions such as Coqui TTS, VITS, and Tacotron.
* Work with multimodal LLMs for different AI use cases.
3) Backend & System Integration
* Build scalable AI APIs using Python, FastAPI, Flask, and REST APIs.
* Integrate AI agents with business systems and telephony platforms such as Asterisk, FreeSWITCH, Twilio, and SIP.
* Ensure reliable real-time audio processing and system integration.
4) Monitoring & Optimization
* Monitor and improve accuracy, response quality, latency, and overall agent performance.
* Implement logging, evaluation, monitoring, and feedback loops.
* Continuously optimize AI agents based on real-world production performance.
๐๐๐ช๐ฎ๐ข๐ซ๐๐ฆ๐๐ง๐ญ๐ฌ:
Bachelorโs or Masterโs degree in CS, AI, ML, or a related field.
3+ years of hands-on experience in AI/ML, NLP, or conversational AI.
* Strong understanding of LLMs, transformers, embeddings, prompt engineering, and fine-tuning.
* Experience with Agentic AI frameworks such as Strands Agents, LangChain, or LangGraph.
* Strong understanding of RAG, vector databases, chunking, retrieval, and context management.
* Experience with multimodal LLMs.
* Proficiency in PyTorch or TensorFlow.
๐๐ฑ๐ฉ๐๐ซ๐ข๐๐ง๐๐ ๐ฐ๐ข๐ญ๐ก:
* STT: Whisper, Wav2Vec2, Kaldi
* TTS: Tacotron, VITS, Coqui TTS
* Backend: FastAPI, Flask, REST APIs
* Telephony: Asterisk, SIP, Twilio, FreeSWITCH
* LLM: Hugging Face, LoRA, PEFT
๐๐จ๐ง๐ฎ๐ฌ ๐๐จ๐ข๐ง๐ญ๐ฌ:
* Experience with Urdu/multilingual NLP.
* Experience with real-time AI voice assistants or call center AI.
* Experience taking AI solutions from prototype to production at scale.
Familiarity with AI monitoring, evaluation, and guardrails.