AI/ML Senior Specialist

JOB DESCRIPTION

Job title: AI/ML Senior Specialist
Location: Kochi (Hybrid)
Contract: Permanent, full-time

About us:

Cyncly is a global technology powerhouse with 2,800+ employees and 70,000+ customers across 100+ countries and 46+ offices. Cyncly transforms the way customizable products and spaces are imagined, designed, sold, managed and made. Our end-to-end software solutions connect professional designers, retailers and manufacturers to the world's largest repository of product content. Today, our business spans across the Kitchen & Bath, Furniture, Window, Glass & Door, and Flooring industries with operations in North & South America, Europe, Asia Pacific and Africa.

Cyncly offers over 30 years of experience to deliver more value for our customers through an expanded portfolio of end-to-end solutions. Our global presence allows us to provide world-class support and sales with a local touch, providing the best possible customer experience.

Cyncly is now embarking on an exciting journey as we continue to expand through strong organic growth and complementary acquisitions, backed by leading growth private equity firms specializing in technology

About the Role

Cyncly is establishing a dedicated AI Centre of Excellence (AI CoE) in Kochi, India, to build, fine-tune, and operationalize on-premises large language models across Cyncly's product suite. The AI/ML Engineer is a hands-on technical role responsible for training pipelines, fine-tuning domain-specific models, embeddings infrastructure, and integrating AI across five LLM use cases: product knowledge search, document intelligence, Service Desk ticket classification and resolution drafting, developer copilot assistance, and conversational product configuration.

Key Responsibilities

LLM Fine-Tuning & Training Pipelines

• Implement and maintain supervised fine-tuning (SFT) and RLHF pipelines for Cyncly domain-specific models, leveraging proprietary product catalogue, documentation, and support data.
• Apply LoRA/QLoRA fine-tuning via Hugging Face PEFT; manage model checkpointing, hyperparameter tuning, and training run tracking. Work with quantised models (GGUF, AWQ, GPTQ) to optimise inference on GPU.
• Maintain reproducible experiments, model versioning, and artefact management using MLflow or Weights & Biases; collaborate with Data Engineering to curate and version domain datasets with PII scrubbing and data-quality checks.

Embeddings Infrastructure & Semantic Search

• Develop and optimise embedding pipelines using models such as BGE-M3 and E5-large; manage vector database ingestion, indexing strategies, and retrieval quality for Cyncly's product catalogues and documentation.
• Design, benchmark, and maintain vector database deployments (Qdrant, Weaviate, or equivalent), including index configuration, filtering strategies, and scalability planning.
• Implement retrieval quality metrics (MRR, NDCG, recall@k), run A/B evaluations of embedding model upgrades, and build hybrid search pipelines combining dense vector retrieval with BM25 sparse search to maximise product discovery accuracy.

RAG Architecture & Use Case Integration

• Design and implement RAG layers for Cyncly's five LLM use cases, ensuring context relevance, faithfulness, and latency targets; integrate MCP connectors to expose internal data sources and product APIs to the LLM inference layer.
• Build and maintain the developer copilot integration using Continue.dev and custom language-server extensions; implement prompt injection defenses, output filtering, and guardrail layers across all inference paths.
• Work closely with Product and Engineering squads to expose AI capabilities through REST/OpenAI-compatible APIs and integrate them into Cyncly's product surfaces.

Model Evaluation & Quality Assurance

• Write model evaluation harnesses and regression test suites; track and report quality metrics (BLEU, ROUGE, faithfulness score, hallucination rate, answer relevance) per deployed model and use case.
• Partner with the AI/ML QA team to design human evaluation workflows and validate model outputs before production release; maintain a model leaderboard and evaluation artefacts for all active models enabling data-driven decisions on upgrades and rollbacks.
• Implement continuous evaluation pipelines triggered on model updates, ensuring quality regressions are detected before reaching production.

Inference Infrastructure & Observability

• Deploy and manage LLM inference using vLLM or Text Generation Inference (TGI) on Cyncly's on-premises GPU infrastructure, targeting production-grade throughput and latency SLAs.
• Instrument inference latency, throughput (tokens/sec), GPU utilisation, and cost metrics; surface results to operational dashboards for the AI leadership team; participate in on-call rotation and incident response for model-serving issues.
• Contribute to the containerisation and orchestration of AI workloads using Docker and Kubernetes, ensuring deployments are reproducible, scalable, and integrated with Cyncly's CI/CD pipelines.

Knowledge Graphs & Advanced Data Pipelines

• Contribute to Cyncly's product knowledge graph (Neo4j) — entity-extraction pipelines, schema design, and graph query integration with RAG; build NLP pipelines to enrich product and document knowledge.
• Explore and prototype agentic AI patterns (tool use, multi-step reasoning) for complex information retrieval and action execution use cases.
• Bachelor's or Master's degree in Computer Science, Artificial Intelligence, Machine Learning, or equivalent practical experience.
• 5–8 years of hands-on software and/or AI/ML engineering experience, with at least 2 years working directly with large language models or transformer-based systems.
• Strong Python programming skills; proficiency with PyTorch or JAX training ecosystems, and Hugging Face Transformers, PEFT (LoRA, QLoRA), and model evaluation frameworks.
• Working knowledge of at least one vector database (Qdrant, Weaviate, Pinecone, Chroma, or equivalent), including indexing, filtering, and retrieval optimisation; solid understanding of embedding models, semantic search quality metrics, and RAG architecture patterns.
• Experience building and consuming REST/OpenAI-compatible APIs; solid understanding of prompt engineering best practices including few-shot prompting, chain-of-thought, and structured outputs.
• Familiarity with containerisation (Docker) and orchestration (Kubernetes), and experience working within CI/CD pipelines.

Mandatory Technical Experience

• LLM Frameworks: Proficiency in Python with the Hugging Face ecosystem (Transformers, Datasets, PEFT, Accelerate, Evaluate).
• Fine-Tuning & Quantisation: Experience with LoRA, QLoRA, and quantisation techniques (GGUF, AWQ, GPTQ) for on-premises model deployment.
• Inference Serving: Hands-on experience with vLLM, Text Generation Inference (TGI), or Ollama for production LLM serving.
• Vector Databases & RAG: Practical experience designing and querying vector databases; proficiency in retrieval evaluation and hybrid search implementations.
• Experiment Tracking & Evaluation: MLflow, Weights & Biases, or LangSmith for experiment tracking and LLM observability; strong background in model evaluation — automated metrics (BLEU, ROUGE, RAGAS), human evaluation, and regression testing.
• Infrastructure & DevOps: Experience with Docker and Kubernetes; ability to write Dockerfiles, Helm charts, and CI/CD pipeline configurations.

Preferred / Nice-to-Have

• Exposure to knowledge graphs (Neo4j, Cypher/SPARQL) for entity extraction and graph-augmented retrieval; experience integrating helpdesk or ticketing platforms (e.g. Zendesk APIs) for NLP-driven classification and resolution drafting.
• Familiarity with agentic AI frameworks (LangChain, LlamaIndex, AutoGen) or MCP connector development; experience with developer tooling integrations such as Continue.dev, GitHub Copilot extensions, or custom LSP implementations.
• Open-source contributions to ML projects, LLM tooling, or evaluation frameworks; experience with multi-modal or vision-language models (VLMs) for product image understanding.

Competency Requirements

• Engineering Excellence: Writes clean, well-tested Python code; comfortable with code reviews, Git workflows, and contributing to shared ML infrastructure libraries.
• Analytical & Experimental Mindset: Designs experiments with clear hypotheses, tracks results rigorously, and uses evaluation data to make evidence-based model decisions.
• Collaboration & Communication: Works effectively with Data Engineers, QA, Product, and Platform teams; communicates ML concepts clearly to non-specialist stakeholders.
• Ownership & Initiative: Able to take ownership of a workstream, manage ambiguity in early-stage AI projects, and deliver iterative results in a fast-moving AI programme.
  

Working for us:

At Cyncly, we call our team OneCyncly, a reflection of how we work together as one, united by our purpose: powering businesses that bring spaces to life. Our strength comes from our diversity of experiences, perspectives, and skills - and we thrive when we work together with openness, trust, and respect.

Here, you’ll join a group of colleagues who take ownership, solve problems, and focus on making an impact. We embrace curiosity, welcome new ideas, and see mistakes as opportunities to learn. You’ll have the freedom to work flexibly and autonomously, supported by teammates and leaders who are committed to your growth.

We celebrate the different ways people contribute and encourage everyone - from every background - to bring their authentic self to work. Because when we collaborate, challenge each other, and share what we know, we build something better together.

If you want to work in a place where your ideas matter, your growth is valued, and your work shapes the spaces people live, work, and play in - come join us.