16 Nov
|
Lumber
|
Bengaluru
AI/ML Engineer (4-7 Years) Company: Lumber, a construction payroll and workforce management platform
Location:
Bengaluru
Experience:
4-7 years in ML/AI, with at least 1-2 years on LLM work
About the Role
We're looking for an AI/ML Engineer who is strong in core machine learning and has hands-on experience fine-tuning and deploying language models in production. You'll build and improve AI-driven features and agents for payroll, compliance, and workforce workflows, where accuracy, reliability, and cost matter.
What You'll Do
● Fine-tune and adapt LLMs (SFT, LoRA/QLoRA, PEFT) for domain-specific tasks such as document understanding, classification, extraction, and reasoning ● Build rigorous evaluation frameworks (offline benchmarks, regression suites, human-in-the-loop review) to measure model quality and catch regressions ● Develop and ship ML models beyond LLMs (classical ML, anomaly detection, ranking, embeddings) where they fit better ● Build RAG pipelines and agentic workflows, including retrieval, embeddings, vector stores, and tool use ● Optimize inference for latency, throughput, and cost (quantization, batching, caching, model selection) ● Deploy and monitor models in production, covering versioning, drift detection,
and observability ● Work with product and engineering teams to turn business problems into ML solutions ● Write explicit technical documentation and mentor junior engineers Required Qualifications ● 4-7 years of experience building and deploying ML systems in production ● Strong foundation in core ML: supervised/unsupervised learning, optimization, regularization, bias-variance, feature engineering, and evaluation metrics ● Solid understanding of deep learning and transformer architectures (attention, tokenization, pretraining vs. fine-tuning) ● Hands-on experience fine-tuning LLMs using PyTorch and Hugging Face (Transformers, PEFT, TRL) ● Experience designing evaluation strategies for both classical models and generative systems ● Strong Python skills and good engineering practices (testing, code review, version control) ● Experience with cloud platforms (AWS preferred) and deploying models via APIs ● Ability to explain trade-offs, such as fine-tuning vs. prompting vs. RAG, to technical and non-technical stakeholders Good to Have ● Experience with self-hosted inference (vLLM, TGI) and managed platforms (AWS Bedrock, SageMaker) ● Familiarity with RLHF/DPO or other preference-optimization methods ● Experience with vector databases (Pinecone, pgvector, etc.) and embedding models ● Knowledge of MLOps tooling (MLflow, Weights & Biases, Docker, CI/CD for models) ● Experience with document AI, OCR, or structured extraction ● Background in fintech, payroll, insurance, or other compliance-heavy domains ● Publications, open-source contributions, or strong applied ML projects
📌 Artificial Intelligence Engineer (Bengaluru)
🏢 Lumber
📍 Bengaluru