Lead Engineer, GenAI

Experience: 4.00 + years

Salary: Confidential (based on experience)

Expected Notice Period: 15 Days

Shift: (GMT+05:30) Asia/Kolkata (IST)

Opportunity Type: Hybrid ()

Placement Type: Full Time Permanent position(Payroll and Compliance to be managed by: MDMY)

(*Note: This is a requirement for one of Uplers client - MDMY)

What do you need for this opportunity

Must have skills required:

AI workflow automation tools, Fine-tuning or RLHF experience, LLM observability tooling, MCP, Multi-tenant AI architecture, Ollama, Tgi, Triton, vLLM, 4+ years building production LLM/GenAI systems, Copilot, LangChain, LangGraph, or autonomous workflow, RAG agent, Strong Python

MDMY is Looking for:

Lead Engineer, GenAI

Location: New Delhi / Gurugram, India

Hybrid (2–3 days on-site)

Function: AI Engineering

Reports to: Director of Data Science and AI

Compensation: Competitive

The Opportunity

Most GenAI roles hand you an API key and ask you to build a chatbot. This one hands you production LLM and VLM systems already running across 15+ enterprise deployments — and asks you to decide what the platform's intelligence becomes next.

We are not at the should we use AI stage. LLMs are live in production: vision models evaluate deployment decisions, extraction pipelines pull structured knowledge from complex engineering documents, a copilot lets teams query operational data through natural language, and LLMs select features for ML models using physics-grounded reasoning. The foundation is built. What is missing is a senior technical owner who sets the direction for how we build AI systems — not just what to ship next.

You are joining as the technical lead of the GenAI sub-team within our largest squad. You will inherit one AI Engineer already building on this stack, and you will grow the team from here. The next layer is copilots that help field engineers diagnose equipment issues in real time, agentic systems that mine years of operational history and propose their own diagnostic rules, and a knowledge infrastructure that makes the entire platform smarter as it scales.

Not prototypes — production systems where the equipment you are building AI for keeps critical infrastructure running. The engineers using your copilots are making maintenance decisions on assets worth hundreds of millions of dollars. You define the ceiling for what the platform's intelligence can do, and you build the team that gets it there.

About Us

We build an enterprise IIoT platform for real-time monitoring, diagnostics, and predictive maintenance of turbomachinery and rotating equipment. Under the hood: a proprietary physics engine (2,689 lines of thermodynamic calculations for 15+ equipment types), custom AutoML, distributed training with Ray, and production AI running on every deployment.

• 45+ employees across New Delhi and Houston
• Venture-funded, closing Series A | 80–90% YoY growth
• 15+ enterprise clients: Chevron, Woodside, Devon Energy, SM Energy, INPEX, Berkshire Hathaway, etc.
• Physics-informed ML + production LLM/VLM: a 12–18 month lead over the field
  
  

What You'll Own


GenAI Sub-Team Leadership (25%) You are the technical lead for the GenAI function:

• Own the technical direction of the GenAI sub-team — architecture decisions, standards, build-vs-buy calls
• Directly manage and grow AI Engineers mentor them toward independent ownership of feature tracks
• Partner with the Director of Data Science and Head of Applied AI on the model strategy roadmap and longer-term AI platform vision
• Define hiring bar for AI Engineers joining your team — you shape who fits and what level we need
• Set the standard for how we build AI systems: eval frameworks, observability, prompt versioning, reliability gates
  
  

Copilot & Conversational AI (30%)

Ship the customer-facing AI products that are the top priority for 2026:

• Own copilot development end-to-end: conversation orchestration, live equipment API integration, historical event context, observability from day one
• Design and build conversational interfaces that operate against real diagnostic rules, sensor data, and equipment metadata
• Work closely with domain engineers to get the reasoning right — your users are enterprise operators making high-stakes decisions
• Define what production-ready means for each system: eval frameworks, regression tests, reliability and latency standards
  
  

Agentic Systems & Platform Intelligence (25%)

Build the loops that give the platform a compounding advantage:

• Design and ship LLM agent workflows that convert anomaly signals into candidate diagnostic rules — human-in-the-loop where needed, automated where earned
• Extend the agentic monitoring roadmap: detection, triage, service request, escalation
• Build internal knowledge infrastructure (RAG + context servers) so every AI tool at Mechademy queries institutional knowledge natively
  
  

Applied AI & Model Roadmap (20%)

• Build and maintain AI-powered automation workflows: integration pipelines, internal skill systems, operational automations
• Contribute to the model roadmap: fine-tuning strategy, dataset curation from proprietary operational data, self-hosted serving architecture, and intelligent routing across a mixed model fleet
• Design eval frameworks for high-stakes AI — how do you measure whether a system diagnosing industrial equipment is correct You define that standard
  
  

What Success Looks Like


First 30 days: Deep familiarity with existing AI pipelines, production systems, and the automation landscape. Working relationship established with your AI Engineer. First architectural decisions made on the copilot stack.

First 60 days: Diagnostic Copilot alpha live with internal domain engineers. GenAI architecture standards documented. Agentic discovery loop design approved. Your AI Engineer shipping independently on an owned feature track.

First 180 days: At least one copilot in customer hands. Agentic automation workflows running across engineering and delivery. Active contribution to the fine-tuning roadmap. Your architectural decisions are the de facto standard for how the team builds. You have a clear view on who the next AI Engineer hire should be.

Your success metric: By end of year, field engineers are using AI your team shipped to make better maintenance decisions, your AI Engineer has grown into owning a full feature track independently, and Mechademy is meaningfully on the path to running its own models.

Who You Are

Must-Have

• 7+ years in ML/AI engineering, with 2+ years focused on LLM/GenAI systems in production — real systems that real users depend on daily, with eval infrastructure and observability to prove it. Not side projects or notebooks.
• Led or technically owned a GenAI product or system — you have made the architectural calls, not just executed on them. You have shipped a customer-facing copilot, RAG agent, or autonomous workflow that users trust with real decisions.
• Agentic framework experience (LangGraph, CrewAI, or equivalent) — you have designed agent state machines, handled tool-calling failures, built human-in-the-loop flows under production conditions. Multi-agent orchestration (orchestrator-specialist patterns, agent memory architectures) strongly preferred.
• LLM evaluation as a discipline — building LLM-as-a-judge rubrics, automated regression suites for AI behavior, and eval gates before deployment. You design evaluations before you write prompts.
• Strong Python — FastAPI/Django integration, async patterns, production-quality code that others can maintain and extend.
• Ability to grow a team — you have mentored junior engineers, reviewed architectural approaches, and contributed meaningfully to hiring decisions.
  
  

Strong Signals

• GraphRAG and knowledge graph architecture — hybrid retrieval combining vector search with graph traversal domain ontology design for asset hierarchies and causal structures
• Context engineering — systematically designing what goes into the LLM context window for agentic tasks running over live data streams. Architectural, not prompt writing.
• Hybrid search architecture (vector + BM25/keyword) and reranking pipelines (cross-encoders, Cohere Rerank, or equivalent)
• Guardrails and safety layer engineering — compliance rails, human-in-the-loop checkpoints, and audit trails for agent systems
• Fine-tuning or RLHF experience — especially on domain-specific datasets
• Self-hosted model serving: vLLM, Ollama, TGI, Triton, or equivalent
• Model routing and multi-model orchestration (cost/latency/capability tradeoffs across a mixed fleet)
• RAG implementation with pgvector, Pinecone, or Weaviate
• LLM observability tooling: Langfuse, LangSmith, or equivalent
• Streaming and real-time AI — running LLM agents over live data streams, not static documents
• Multi-tenant AI architecture — per-client data isolation in AI systems
• Experience with AI workflow automation tools (n8n, LangFlow, or similar)
• Time-series or sensor data background — if you have this AND strong LLM experience, you are exceptionally rare in this market
• IoT, energy, or industrial sector domain experience — not required, but it compresses your ramp significantly
  
  

The Right Mindset

• Leads, not just delivers — you are accountable for what the team ships, not just your own output. You feel the difference.
• Ships first — bias for working and in users hands over perfect and in a branch.
• Eval-obsessed — the only AI systems you trust are the ones you have measured.
• Domain-curious — you find it interesting that the features in our ML models were derived from real physics and engineering principles. You do not need to know the domain already — the domain is learnable, and you will pick up industry context faster here than anywhere else.
• Comfortable with ambiguity — some of what you will build has no playbook. You can design from first principles.
  
  

Why Us


The infrastructure is real. You're not joining to add AI to a legacy product. LLMs are running in production on every model deployment. LangGraph is live. You're building the next layer, not the first.

The model roadmap is genuinely interesting. Most applied AI roles cap at call the API better. Here, you will work on the full arc: enterprise LLMs today, fine-tuned small language models on proprietary industrial data, self-hosted model infrastructure for enterprise data privacy, intelligent routing across a mixed model fleet.

The data is irreplaceable. Years of expert diagnostic histories, real operational sensor data across 15+ enterprise clients, and physics-grounded feature sets. This data does not exist in any public dataset. When you fine-tune a model on this data, it learns domain knowledge no foundation model has ever seen.

The portfolio is career-defining. Production copilots shipped, agentic systems running, a team you have grown, and technical leadership you own. Most AI engineers take years to accumulate this breadth — and fewer still get the lead role alongside it.

The team is honest. We have built serious AI infrastructure by learning hard things — and we know exactly what is missing: someone with the applied AI craft and leadership capability to take it further. You are not joining a lab. You are joining a team that will learn from you.

What We Offer

• Production AI infrastructure from day one: enterprise LLMs, LangGraph, Langfuse, Ray, MLflow, pgvector
• A genuine model roadmap: hosted APIs, fine-tuned SLMs, self-hosted inference — you shape it
• A team to lead from day one: one strong AI Engineer ready to grow under you, more headcount coming
• Close collaboration with domain engineers who carry deep equipment knowledge
• Real feedback loop: internal team and engineers at enterprise clients using your AI daily
• Hybrid flexibility: 2–3 days on-site in Gurugram, rest remote
• Competitive compensation
  
  

Qualifications

• B.Tech / B.E. / B.S. in Computer Science, Engineering, Mathematics, or a related discipline — or equivalent experience
• 7+ years of ML/AI engineering experience, with 2+ years on LLM/GenAI systems
• Master's preferred but track record matters more than credentials — show us what you have shipped and what you have led
• Startup or high-growth environment experience strongly preferred
  
  

How to apply for this opportunity

• Step 1: Click On Apply! And Register or Login on our portal.
• Step 2: Complete the Screening Form & Upload updated Resume
• Step 3: Increase your chances to get shortlisted & meet the client for the Interview!
  
  

About Uplers:


Our goal is to make hiring reliable, simple, and fast. Our role will be to help all our talents find and apply for relevant contractual onsite opportunities and progress in their career. We will support any grievances or challenges you may face during the engagement.

(Note: There are many more opportunities apart from this on the portal. Depending on the assessments you clear, you can apply for them as well).

So, if you are ready for a new challenge, a great work environment, and an opportunity to take your career to the next level, don't hesitate to apply today. We are waiting for you!