Senior AI/DS Engineer

Designation : Senior AI/DS Engineer

Required Experience: 5-6 years overall with AI/ML, data science or applied AI; at least 5 years of hands-on AI/ML/data-science work and 2+ years of technical/team leadership.

Job Description:

Aventior is looking for a hands-on Team Lead - AI & Data Science to lead customer-facing AI and analytics engagements from discovery through production and ongoing improvement.

The role combines technical direction, people leadership, customer communication and delivery ownership. You will turn business problems into clear requirements, assess feasibility and data readiness, shape pragmatic AI/GenAI solutions, guide engineers and data scientists, and deliver secure, measurable outcomes in production.

The strongest candidate is equally comfortable discussing business value with a customer, explaining technical options and trade-offs, reviewing code or model quality, and managing scope, risks and delivery commitments.

Key Responsibilities:

1. Customer engagement, discovery and requirement gathering -

●Lead customer and stakeholder discovery sessions, interviews and workshops to understand business goals, users, existing workflows, pain points, data sources, constraints and success criteria.

●Translate business needs into clear problem statements, functional and non-functional requirements, user stories, acceptance criteria, assumptions, dependencies and measurable KPIs.

● Assess data availability, data quality, integration constraints, privacy/security requirements and technical feasibility before recommending an approach.

● Present solution options and trade-offs in business-friendly language, including build-versus-buy, model selection, expected value, risks, cost, latency and operational implications.

● Lead proof-of-concept planning, demos, review meetings and customer sign-off; manage clarifications, scope changes, priorities and expectation setting throughout the engagement.

2. AI solution architecture and technical leadership

● Own solution design across data ingestion and preparation, model or LLM selection, retrieval, orchestration, APIs, integrations, cloud infrastructure and operational monitoring.

● Choose fit-for-purpose approaches for the business problem rather than defaulting to a particular model, vendor, framework or architecture pattern.

● Establish engineering standards through design reviews, code reviews, testing practices, reusable components, documentation and secure development practices.

● Remain hands-on for critical design decisions, prototypes, complex debugging and performance/cost optimisation while enabling the team to execute independently.

3. Generative AI, LLMs, RAG and agentic workflows

● Design and deliver LLM-powered applications using commercial APIs and/or managed or open models; apply prompt design, structured outputs, tool/function calling and workflow orchestration as appropriate.

● Build and improve RAG pipelines, including document ingestion, chunking, embeddings, vector or hybrid retrieval, reranking, access-aware retrieval, source attribution and grounded answer generation.

● Select and apply agentic patterns only where they add value; define tool permissions, bounded actions, state transitions, approval steps, retries and safe failure behavior.

● Create systematic evaluation approaches for task success, groundedness, retrieval relevance, factuality, safety and user experience; use representative test sets, regression tests and human review where needed.

●Improve quality, latency and cost through retrieval/model comparisons, prompt and context tuning, caching, routing, token-usage controls and production feedback loops.

4. Applied machine learning and data science

● Guide the framing, development and validation of appropriate ML/statistical solutions such as classification, regression, forecasting, anomaly detection, clustering, NLP or optimization.

● Establish baselines and evaluation criteria; interpret results, investigate failure modes, check for data leakage and communicate limitations and confidence clearly.

● Partner with data engineers and analysts on data profiling, feature engineering, data pipelines, dataset quality, reproducibility and reliable access to data.

● Guide specialist computer-vision, OCR/intelligent document processing or multimodal work when required by the project portfolio; deep specialist research methods are not a blanket requirement for this role.

5. Production engineering, cloud and MLOps / LLMOps

●Build or guide production services and integrations using Python and REST APIs (FastAPI or equivalent), with appropriate validation, authentication, error handling, testing and documentation.

●Deploy and operate solutions in AWS or the agreed cloud platform. Relevant AWS services may include S3, SageMaker, Bedrock, RDS, Glue, EventBridge, Redshift, Lambda and container/orchestration services, depending on the architecture.

●Implement or oversee CI/CD, containerization, environment management, model/prompt/version tracking, automated tests, release controls, rollback and operational handover.

●Monitor application health and business/model quality, including latency, throughput, availability, cost, token consumption, retrieval quality, feedback signals and model/data drift where applicable.

● Own incident triage and root-cause analysis with engineering/platform teams; define fallback paths and support processes for customer-facing AI capabilities.

6. Delivery management and stakeholder alignment

●Own the delivery plan from discovery to launch: estimates, milestones, sprint/backlog priorities, resource needs, dependencies, risks, testing, UAT, release readiness and transition to support.

●Maintain a clear decision, action and risk log; provide concise progress updates; escalate issues early and propose options instead of only reporting blockers.

●Coordinate across customer teams, product/business analysts, data engineering, software engineering, QA, cloud/DevOps, security and domain experts.

●Ensure acceptance criteria and agreed business outcomes are met, and capture lessons learned and post-launch improvements

7. Team leadership, governance and responsible AI

● Coach and mentor AI engineers and data scientists; support work allocation, technical growth, feedback, code/model reviews and consistent quality standards.

●Apply appropriate data governance, privacy, access control, PII handling, audit logging, human review and responsible-AI safeguards based on use-case risk.

●Identify GenAI-specific risks such as prompt injection, insecure tool access, unintended data exposure, unsupported outputs and misuse; document mitigations and residual risks.

●Build reusable patterns, documentation and team practices that improve delivery predictability across multiple customer engagements

Required Qualifications and experience:

●5-6 years of relevant software engineering, AI/ML, data science or applied

AI experience, including at least 5 years of hands-on delivery experience.

●At least 2 years leading a technical team, delivery workstream or project; demonstrated ownership of technical decisions and outcomes.

●At least 2 years in customer-facing or stakeholder-facing delivery, with evidence of running discovery workshops, clarifying requirements, presenting options and managing expectations.

●Strong Python engineering and practical SQL/data skills; understanding of testing, APIs, code quality and software engineering fundamentals.

●Strong applied ML/statistical reasoning and model evaluation skills, including the ability to choose baselines, validate results and explain limitations.

●At least 2 years building GenAI/LLM applications, including practical RAG experience; able to explain architecture, evaluation, failure modes and production trade-offs. At least one production deployment is strongly preferred.

●Working experience with AWS or the target cloud platform, and with model/application deployment, CI/CD, observability and production support.

●Experience with REST APIs using FastAPI or an equivalent Python web framework, and with at least one vector database or semantic-search technology when RAG is in scope.

●Strong communication, facilitation, documentation, mentoring and problem-solving skills.

●Hands-on experience with agentic/tool-using workflows, multimodal LLMs, VLMs, or more than one model provider.

●Computer vision, OCR, intelligent document processing, object detection, image segmentation or document-understanding experience where relevant to the client pipeline.

●Experience with model fine-tuning, embedding adaptation, open-source model serving or GPU/inference optimization.

●Familiarity with evaluation and tracing tools such as LangSmith, OpenTelemetry, Phoenix or equivalent; equivalent tools are acceptable.

●Experience in consulting, professional services or multi-client delivery environments.

●Bachelor's or Master's degree in Computer Science, Engineering, Data Science, Statistics or a related field. Equivalent practical experience may be considered.

Company Website: https://aventior.com/