[Remote] LLM Fine-Tuning Engineer

Note: The job is a remote job and is open to candidates in USA. Bright Vision Technologies is a forward-thinking software development company dedicated to building innovative solutions that help businesses automate and optimize their operations. They are seeking an LLM Fine-Tuning Engineer to design, execute, and operationalize fine-tuning workflows for large language models, collaborating with cross-functional teams to translate requirements into well-engineered solutions. Responsibilities Design and execute fine-tuning experiments for large language models using supervised, DPO, RLHF, and related techniquesLead dataset construction, curation, and quality assurance processes for instruction tuning and preference dataBuild scalable training pipelines on top of modern distributed training frameworksTune hyperparameters, optimizer configurations, and training stability strategies for large-model fine-tuningImplement parameter-efficient fine-tuning techniques such as LoRA, QLoRA, and adapter-based methodsDesign rigorous evaluation suites including automated benchmarks, human evaluation, and capability-specific probesImplement safety, refusal, and policy evaluations to track model behavior across releasesOperate large-scale training jobs on GPU clusters, diagnosing failures and recovering training state reliablyOptimize training throughput using mixed precision, sequence packing, and efficient attention implementationsManage model artifacts, lineage tracking, and reproducibility across many concurrent experimentsCollaborate with product, research, and platform teams to align fine-tuning roadmaps with business needsDocument training methodology, results, and decisions clearly for technical and non-technical audiencesMentor engineers on fine-tuning best practices, evaluation rigor, and responsible deploymentStay current with LLM research and translate advances into production-ready fine-tuning recipes Skills Master's or PhD in Computer Science, Machine Learning, or a related field; or equivalent experienceSix or more years of combined ML research and engineering experience, with significant LLM exposureStrong proficiency in Python and modern deep learning frameworks, especially PyTorchHands-on experience fine-tuning transformer-based language models at non-trivial scaleFamiliarity with distributed training strategies including FSDP, ZeRO, and pipeline parallelismExperience with RLHF, DPO, or other preference optimization techniquesStrong understanding of evaluation methodology, benchmarks, and human evaluation designExperience operating training jobs on GPU clusters and recovering from failuresStrong written and verbal communication skillsTrack record of shipping or publishing impactful LLM workPublications at top-tier ML venuesExperience with multimodal model fine-tuningFamiliarity with synthetic data generation and dataset distillationOpen-source contributions to LLM training librariesExposure to responsible AI evaluation and red-teaming practices Benefits Full-time, direct W2 with Bright Vision Technologies (no C2C, no 1099, no third-party)Competitive base salary commensurate with experience, plus benefits.We will support H1B transfers for qualified candidates. Company Overview Bright Vision Technologies is an information technology company that offers software development, AI, and cybersecurity services. It was founded in 2020, and is headquartered in Bridgewater, New Jersey, USA, with a workforce of 51-200 employees. Its website is https://bvteck.com.