DeveloperJobs.io
← Back to all jobs

Job Description

The Bison Group LLC is building AI analysis tools that rely on robust model training and careful evaluation from data preparation through deployment. In this onsite role in El Paso, TX, you will own key parts of the model training lifecycle for open-weight language and retrieval models, with additional hands-on work across RAG and speech-related models.

You will adapt models to domain-specific language and tasks, then validate improvements with offline evaluation and regression testing before packaging and serving on-premises on constrained GPU and hardware environments.

What you will do

  • Train and fine-tune open-weight language models using supervised fine-tuning and parameter-efficient methods such as LoRA and QLoRA.
  • Improve model performance on domain terminology, task instructions, and structured outputs.
  • Build instruction and preference datasets from transcripts, surveys, and reference materials.
  • Clean, deduplicate, and label data, and generate and validate synthetic examples with subject matter support.
  • Train or fine-tune embedding and reranker models for retrieval-augmented generation (RAG).
  • Measure retrieval quality for specialized terminology and acronyms.
  • Adapt speech recognition, speaker diarization, and voice activity detection models for noisy audio, varied accents, and multi-speaker conversations.
  • Build offline evaluation suites and regression tests covering citation and attribution accuracy, hallucination rate, and latency.
  • Compare results against established baselines prior to quantization, export, and serving of trained models on on-premises GPUs and constrained hardware.
  • Balance model quality with memory use, inference speed, and throughput.
  • Create reproducible training workflows using experiment tracking, versioned datasets, checkpoints, and model cards.
  • Package models and dependencies for deployment in environments without internet.
  • Partner with product, UX, analysts, and subject matter experts to define useful model behavior.
  • Document experiments and decisions so teams can retrain, evaluate, and maintain models.

What you bring

  • 3+ years of hands-on neural network training or fine-tuning.
  • Ability to explain projects end to end, including data preparation, training decisions, evaluation results, and deployment.
  • Hands-on experience fine-tuning a 7B+ parameter open-weight language model using SFT, LoRA/QLoRA, or full fine-tuning, with PyTorch and Hugging Face Transformers/PEFT, NeMo, or equivalent tools.
  • Strong dataset construction skills: cleaning, deduplication, formatting, labeling, and assessing whether data is suitable for training.
  • Experience designing task-specific metrics and test sets, and diagnosing overfitting, catastrophic forgetting, and data leakage.
  • GPU training fundamentals including mixed precision, gradient accumulation, checkpointing, memory budgeting, and multi-GPU basics.
  • Strong Python and Linux skills, including command-line work, CUDA environment management, and dependency management without internet access.
  • U.S. citizenship and eligibility to obtain and maintain a U.S. security clearance.

Technologies

  • Python, Linux, CUDA
  • PyTorch, Hugging Face Transformers, PEFT, NeMo
  • LoRA, QLoRA, supervised fine-tuning (SFT)
  • Embedding models, reranker models, retrieval-augmented generation (RAG)
  • Speech recognition, speaker diarization, voice activity detection
  • Quantize, experiment tracking, model cards
  • On-premises GPUs, mixed precision, gradient accumulation, checkpointing, multi-GPU basics

Preferred qualifications

  • iOS development, especially on-device machine learning with Core ML or MLX.
  • Training small language models or distilling larger models for edge and on-device use.
  • Experience with NVIDIA NeMo or model serving with NIM/TensorRT-LLM.
  • Preference tuning with DPO, ORPO, or RLHF, or experience with reward modeling.
  • Training embedding or reranker models for RAG applications.
  • Speech recognition fine-tuning, speaker diarization, or voice activity detection.
  • Quantization and inference tooling such as GGUF, llama.cpp, vLLM, AWQ/GPTQ, or MLX.
  • Experience with ARM64 or other unified-memory platforms.
  • Work in defense, intelligence, or air-gapped environments; an active Secret clearance or higher.
  • Familiarity with qualitative research data, including focus groups, surveys, and interviews.

Similar Jobs