Overview

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

Responsibilities

• 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.
• 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.
• 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.
• 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.
• 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.
• 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.
• Knowledge Graphs & Advanced Data Pipelines

Qualifications

• Familiarity with containerisation (Docker) and orchestration (Kubernetes), and experience working within CI/CD pipelines.
• 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.
• 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.
• 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.