Data Scientist III

Rekruton Global IT Services; Ahmedabad District; On-site; INR 3,000,000 - 5,500,000; Full time

Job summary

Rekruton Global IT Services seeks a Data Scientist III to own the analytical strategy for a core product area, delivering a production-grade, monitored model system. You will architect data pipelines, ETLs, and model-serving workflows, while mentoring peers.

You will apply LLMs, prompt engineering, and RAG to real product problems, ensure proper governance, and actively contribute to scalable AI solutions across the team.

Qualifications

• Education in Computer Science, Data Science, or related field.
• 5+ years hands-on data science & ML experience.
• Hands-on MLOps across training, deployment, monitoring.
• Experience building scalable ML models for production.
• Experience with LLMs, RAG, embeddings, and vector databases.
• Strong SQL knowledge and data mindset.
• Experience fine-tuning generative AI models is a plus.

Responsibilities

• Own end-to-end analytical framework from problem framing to production monitoring.
• Apply LLMs, prompt engineering, and RAG to product problems.
• Evaluate when AI is appropriate vs traditional models.
• Mentor junior data scientists and review their work.
• Design experiments (A/B tests, causal inference, uplift).
• Drive adoption of MLOps tooling and scalable modeling practices.

Skills

MLOps

LLMs

Prompt engineering

Vector databases

LangChain

A/B testing

SQL

Model deployment

Education

B.Tech / MS in CS/Data Science

Tools

Pinecone

FAISS

Weaviate

LangGraph

AWS Bedrock

Job description

Job Role

As a Data Scientist III, you will own the analytical strategy for a core product area and take it all the way to a governed, monitored production system at scale, not just a working notebook. This role owns model-serving workflows end-to-end: you will define and optimize the data pipelines and ETLs that feed your models, build models that hold up under real production load, and implement the monitoring and governance that keeps them trustworthy over time. You will also mentor other data scientists doing the same.

Key Responsibilities

• Own the end-to-end analytical framework for a product area, i.e., from problem framing through production deployment and monitoring, with minimal oversight
• Apply LLMs, prompt engineering, and retrieval-augmented generation (RAG) to real product problems — e.g., translating complex targeting/attribution outputs into plain language insights, or building AI-assisted workflows for internal or client-facing tools.
• Evaluate when generative AI is the right tool versus when a simpler model or heuristic will outperform it. This role is expected to have informed judgment here, not default-to-LLM instincts
• Stay hands-on: personally build, code, and develop AI/ML solutions rather than only architecting and delegating
• Build models that scale to production data volumes and real-time serving loads, not just prototypes that work on a sample
• Define and optimize data pipelines, ETLs, and AI model-serving workflows for your domain; this is not handed off to a separate MLOps function
• Implement model monitoring and governance (drift detection, performance tracking, retraining triggers, audit trails) so that production models stay reliable and defensible under compliance scrutiny
• Design measurement approaches for problems where clean ground truth doesn't exist
• Mentor and provide technical guidance to Data Scientist I/II team members; review and approve their modeling and code before it ships.
• Design and defend experimentation approaches (A/B testing, causal inference, uplift modeling) that hold up under stakeholder and compliance scrutiny
• Drive adoption of better MLOps tooling, scalability practices, or modeling methods across the data science team
• Proactively identify whitespace opportunities in product roadmap and bring data-backed recommendations to stakeholders, not just answers to their questions.

Skills Required

• Education: B.Tech / MS in Computer Science, Data Science, or a related field.
• At least 5 years of hands-on experience in data science and machine learning (both) is a must have
• Hands-on MLOps experience is a must have. You should be comfortable owning a model from training through deployment, serving, and production monitoring, not handing it off once it works in a notebook. This includes experience with model-serving infrastructure, pipeline/ETL orchestration, and monitoring or governance tooling (drift detection, performance dashboards, retraining triggers)
• Proven experience building and scaling machine learning models to handle production data volumes, throughput, and latency requirements is a must have. This means designing for scalability up front, not retrofitting it after something breaks in production
• Hands-on experience with LLMs and RAG architectures is a must have. This includes practical experience with prompt engineering, embedding-based retrieval, and vector databases (e.g., Pinecone, FAISS, Weaviate), and frameworks like LangChain/LangGraph or AWS Bedrock
• Demonstrated ownership of at least one modeling project taken fully end-to-end into production, with measurable business impact
• Experience designing and running experiments (A/B testing, or causal inference methods such as DiD, IV, PSM, uplift modeling) is non-negotiable
• Track record of mentoring or reviewing the work of junior data scientists is a must have
• A strong knowledge of SQL, data structures, and query optimization at a level where you can review others' queries, not just write your own
• Experience fine-tuning generative AI models (not just calling APIs) for real-world applications is highly desirable
• Familiarity with AI agent orchestration and multi-step agentic workflows is a plus
• Prior experience in advertising technology, programmatic media buying, or another high-scrutiny/regulated data domain is highly advantageous and good to have.

Personality & Work Ethic

• A strategic thinker who sets technical direction for a problem area, not just aligns to one
• Comfortable operating in ambiguity where the "right" ground truth doesn't exist yet
• Excellent communication skills, with a track record of influencing product or technical decisions using data
• Genuine mentorship instinct; has grown at least one junior team member's skills, not just supervised their output
• Ownership mentality that extends past "the model works" to "the model is scalable, monitored, governed, and still trustworthy six months from now."
• Doesn't let a research rabbit hole delay a hard deadline

Skills

etl pipelines,machine learning,faiss,llm,model training,data modelling,pinecone,langchain,mlops,a/b testing,data science,langgraph