Lead Architect

Merkle Italia; Karnataka; Hybrid; INR 4,000,000 - 7,000,000; Full time

Job summary

Merkle Italia in Bengaluru, India, seeks an AI Solutions Lead to architect and govern our AI delivery practice across GenAI, Agentic AI, and applied ML engagements. This hybrid role combines hands-on solution design with strategic leadership of the AI craft and team.

You will drive reference architectures, evaluation standards, and reusable components while mentoring engineers and collaborating with DX and engineering leadership to raise the technical bar across client engagements.

Qualifications

• 10–12 years hands-on experience building and deploying ML/DL systems in production.
• 3+ years leading ML-AI pods or teams, mentoring engineers and setting standards.
• Experience with enterprise-grade deployments, cost and latency considerations.

Responsibilities

• Translate client business problems into staged, defensible AI solution roadmaps.
• Lead solutioning and provide architecture for end-to-end GenAI, Agentic AI, multimodal and ML use cases.
• Own reference architectures and design patterns for multimodal agentic systems and tools.
• Design and lead multi-agent systems with memory, grounding, and inter-agent communication patterns.
• Establish evaluation frameworks, guardrails, and production quality for AI systems.

Skills

AI Architecting

ML/DL fundamentals

GenAI/Agentic AI

Solution design

Leadership

Tools

LangGraph

LlamaIndex

AutoGen

Knowledge graphs

Job description

Job Description: AI Solutions Lead/AI (Associate) Architect Role Overview

We are seeking an AI Solutions Lead to architect, govern, and grow our AI delivery practice across GenAI, Agentic AI, and applied ML engagements. This is a hands‑on and hybrid role that includes architecting AI solutions and shaping the growth of the AI practice. This role is suited for someone who has earned technical fluency with GenAI and agentic AI on top of a strong foundation in classical ML and deep learning - and who is now ready to set the technical direction for a growing team. The role is a hybrid lead position: leading solutions for multiple AI delivery pods, partnering with engineering and DX leadership, and owning the craft of the practice - driving solution architectures, evaluation standards, reusable components, and the technical bar for the AI delivery team. The expectation is depth and breadth showcased across the portfolio of AI work done so far, preferably in multimodal agentic systems, SLM design, ML and DL solutions, and enterprise deployments, while consistently raising the standards at which the team operates.

Key Responsibilities

Solution Architecture & Technical Direction Translate business problems from clients into staged, defensible AI solution roadmaps working with business leaders through pre-sales and project delivery cycles. Lead solutioning, support architecture for end-to-end AI solutions across GenAI, Agentic AI, multimodal, and applied ML use cases, with explicit trade‑off analysis on model class (frontier vs. SLM vs. fine‑tuned), retrieval design, memory, and orchestration. Own the practice's reference architectures and solution design patterns for multimodal agentic systems, including planning, tool use, memory, grounding, and inter‑agent communication (MCP, A2A). Conduct solution design reviews across concurrent client engagements; facilitate subjective technical decisions and enable delivery excellence. Multimodal Agentic Systems & SLM Design Design and lead the build of multi-agent systems with reasoning, planning, tool use, persistent memory, and grounded retrieval. Guide multimodal system design across text, vision, speech, and structured data, including ingestion, representation, and downstream agent reasoning. Establish patterns for SLM design and adoption - distillation, fine‑tuning, quantization, and routing - to meet enterprise constraints on cost, latency, data residency, and on‑prem/edge deployment. Define hybrid retrieval and knowledge architectures spanning vector, graph (KG), and NoSQL stores; lead KG-assisted retrieval, entity linking, and structured grounding. Eval, Guardrails & Production Quality Establish evaluation as a first‑class discipline: design eval frameworks, golden datasets, regression suites, automated and human‑in‑the‑loop evals, and observability for agentic and generative systems. Define and enforce safety, guardrail, and hallucination‑control standards across the practice; lead red‑teaming and adversarial testing for high‑stakes deployments. Set the bar for production readiness - reliability, latency, cost, monitoring, drift detection, and incident response - for AI systems in regulated, enterprise-grade environments. Drive enterprise deployment best practices across cloud hyper‑scalers, on‑prem, and edge, including GPU/accelerator ops, model serving, and lifecycle automation. Practice Building & Technical Mentorship - Shape the practice's capability roadmap: which techniques to invest in, which to retire, and how the team stays at the leading edge of GenAI and agentic AI. Mentor AI Engineers and Lead AI Engineers; run technical reviews, pairing sessions, and internal knowledge exchange on agentic, multimodal, and SLM topics. Set the technical hiring bar; lead architecture and senior engineering interviews and calibrate the team's evaluation standards. Establish and promote AI in SDLC frameworks on delivery projects Cross-functional Leadership & Delivery Partner with engineering, data science, product, and DX leadership on delivery and acceleration initiatives Engage with client and stakeholder leadership on architecture, feasibility, and risk; communicate technical direction clearly to non‑technical audiences. Support pre-sales and solutioning for new GenAI and Agentic AI opportunities, including effort estimation, architectural framing, and capability storytelling.

Required Technical Skills – Programming & Engineering

Python (advanced), SQL; strong API and backend engineering in FastAPI/Flask/Django; production‑grade software practices. Generative AI: LLMs and SLMs, RAG/Agentic RAG, multimodal architectures, agents, prompt engineering, grounding, knowledge graphs, fine‑tuning (SFT, LoRA/QLoRA, RLHF/RLAIF), distillation, and quantization. Agentic AI: Multi‑agent orchestration, planning, tool use, persistent memory, MCP and A2A patterns; frameworks such as LangGraph, LlamaIndex, AutoGen. Eval & Safety: Eval framework design, golden datasets, automated and human evals, red‑teaming, guardrails, hallucination control, observability for AI systems. Machine Learning & Deep Learning: Predictive modeling, deep learning (CNNs, RNNs/LSTMs, Transformers), embeddings, vector search, classical ML; CV, NLP, and time‑series exposure. Cloud, MLOps & Deployment: AWS, Azure, or GCP at depth; model serving, GPU/accelerator ops, CI/CD, monitoring, on‑prem and edge deployment patterns. Data Engineering: Kafka, Spark/Flink, Hadoop, MongoDB and other NoSQL/graph/vector stores; large‑scale streaming and batch pipelines. Math Foundations: Linear algebra, probability, statistics, optimization. Experience with commerce cloud ecosystems (good to have) - Salesforce and Adobe Experience Requirements - 10-12 years of hands‑on experience building and deploying ML, DL, and AI systems in production, with progression into solution architecture and technical leadership 10+ years of demonstrable experience working with global businesses, delivering on large accounts 3+ years of demonstrable hands‑on work in GenAI and/or Agentic AI - beyond prompt engineering and basic RAG - including multi‑agent systems, custom fine‑tuning, multimodal pipelines, or SLM-based deployments. Proven track record of architecting and shipping AI systems in enterprise‑grade environments, including regulated or high‑stakes domains. 3+ Experience leading ML‑AI technical pods or teams (formal or dotted‑line), mentoring senior engineers, and setting hiring and review standards.

Attitude & Mindset

You must have an architect's mindset and equipped with a builder's hands. You must be agile and current - actively in the thick of GenAI and agentic developments, learning and shipping at the pace the field demands, with strong fundamentals and understanding of the domain. Excellent communicator - able to explain agentic, multimodal, and SLM trade‑offs clearly to engineering peers, business stakeholders, and clients. Open and flexible toward a hybrid work structure with no less than 3 days work from office - to ensure regular connection and cross‑project knowledge exchange across the AI practice.

Skills Matrix

Skill Category

• AI Solutions Lead Solution Architecture Owns reference architectures across multiple concurrent AI engagements; sets patterns for agentic, multimodal, and SLM‑based systems.
• Transformers & Deep Learning Deep practical grounding in transformers, fine‑tuning (SFT, LoRA/QLoRA, RLHF/RLAIF), distillation, quantization, and SLM design for cost/latency‑bound deployments.
• Generative AI (LLMs & Multimodal) Designs hybrid multimodal RAG, KAG, and grounded‑generation systems; selects the right model class (frontier vs. SLM vs. fine‑tuned) per use case.
• Agentic Frameworks Architects multi‑agent orchestration, planning, memory, tool use, and inter‑agent communication patterns including MCP and A2A.
• Eval, Safety & Guardrails Establishes evaluation frameworks, hallucination control, red‑teaming, regression suites, and observability as a first‑class engineering discipline.
• Information Retrieval & Knowledge Designs hybrid dense‑sparse retrieval, ranking, KG‑augmented retrieval, and unified knowledge layers across vector, graph, and NoSQL stores.
• Predictive & Classical ML Strong foundation in classical ML and DL (CV, NLP, time‑series, GNNs); able to choose non‑GenAI approaches when they fit better.
• Conversational AI Architects multi‑turn, multilingual, multimodal dialogue systems with grounded responses and structured evaluation.
• Model Deployment Drives enterprise deployment patterns - cloud, on‑prem, and edge - including GPU/accelerator ops, scaling, CI/CD, and lifecycle automation.
• Cloud & MLOps End‑to‑end model and agent lifecycle on AWS/Azure/GCP; cost, latency, and reliability optimization at production scale.
• Data Engineering & Pipelines Designs streaming and batch pipelines (Kafka, Spark, Flink) and the data foundations that production AI systems depend on.
• Practice Leadership Sets the technical bar for hiring, mentoring, code/architecture reviews, and reusable IP; grows the AI practice as a craft.

Location: DGS India - Bengaluru - Manyata H2 block Brand: Merkle Time Type: Full time Contract Type: Permanent

About dentsu

For over 120 years, innovation has been a core tenet of our offering - exploring new ways to reach, engage and nurture relationships with audiences. Together we drive a multiplier effect …