About the role
Clients are asking how to turn scattered data and AI curiosity into real outcomes. You'll build the platforms underneath, and deliver applied AI that is governed, evaluated and measurably useful.
Expect to move between architecture, hands-on engineering and explaining trade-offs to business stakeholders.
Your team
You'll work with data engineers, analysts and software engineers, and share an evaluation framework used across AI engagements.
The impact you'll have
Trusted data
Governed, well-modelled data products that teams rely on.
AI that ships
RAG and agent solutions with evaluation, security and cost controls.
Decisions faster
Analytics that people use because it answers their questions.
What you'll do
- Architect lakehouse, streaming and analytics platforms
- Build governed RAG and agent solutions with evaluation and guardrails
- Translate business questions into data products
- Establish MLOps and data quality practices
A typical week might include
- Modelling data in dbt
- Evaluating an LLM pipeline against a test set
- Workshop with business stakeholders
- Reviewing data quality alerts
Technical scope and expected depth
The areas this role may cover. The exact environment and priorities are agreed for each engagement.
Data architecture and pipelines
Assess ingestion, transformation, storage, orchestration, and consumption paths. Define ownership, data contracts, freshness expectations, and recovery behavior.
Quality and governance
Create practical checks for completeness, consistency, accuracy, lineage, and access. Explain how a known data limitation changes the decisions a consumer can safely make.
Analytics and semantics
Translate business questions into shared metric definitions, dimensional models, and analytical experiences. Reconcile competing definitions before building a new dashboard.
Applied AI workflows
Prototype retrieval, summarization, classification, or assisted decision workflows where justified. Assess context quality, model behavior, workflow fit, and the need for human approval.
Evaluation and monitoring
Define representative evaluation examples, acceptance criteria, error analysis, and production monitoring. Track quality, latency, cost, and user adoption together.
Access and responsible operation
Plan sensitive-data handling, permissions, retention, auditability, and fallback behavior with security and business owners before scaling access.
What you'll deliver
Decision map
A shared view of the decision, user, workflow, signal, consequence, and business outcome the work should improve.
Foundation plan
Priorities for data quality, architecture, governance, access, ownership, and the operating model behind trustworthy information.
Working prototype
A focused experiment that makes the opportunity tangible and exposes the real product, data, and adoption questions.
Responsible scale path
Evaluation, controls, measurement, change enablement, and sequencing for moving from promising to dependable.
What you bring
- 5+ years in data engineering, analytics or ML
- Strong SQL, Python and a modern warehouse (Snowflake, Databricks, BigQuery)
- Experience shipping LLM-based features to production
- Comfortable presenting to non-technical stakeholders
Practical capabilities we will discuss
- Data engineering fluency
- Explain SQL, data modeling, transformation, orchestration, and analytical consumption through a real project. Discuss how freshness, scale, access, and failure recovery influenced your design.
- AI evaluation
- Describe how you would create a representative evaluation set, classify failure modes, compare alternatives, and establish acceptance criteria for a proposed AI workflow.
- Product and adoption
- Show how you connect a data or AI output to a human decision. Explain the workflow change, review responsibility, quality threshold, and feedback required for adoption.
- Evidence of impact
- Bring an example in which a data product, analytical capability, or AI experiment changed an operational decision. Explain what you measured, what remained uncertain, and why the next investment was justified.
Nice to have
- dbt, Airflow, vector databases
- Responsible-AI and privacy experience
How success is assessed
Success criteria are agreed at the start of the engagement and reviewed as the work progresses. Typical signals include:
- The organization can name the decision and user the data or AI capability is meant to support.
- Data quality, ownership, access, and limitations are visible enough to guide responsible use.
- Prototypes are evaluated against value, adoption, risk, and operational cost—not excitement alone.
- Leaders can choose the next investment with more confidence and less vendor-driven ambiguity.
- The client team gains a repeatable way to identify, test, govern, and scale useful intelligence.
Your first 90 days
First 30 days
- Learn a client's data landscape
- Contribute to pipeline or model work
- Review our AI evaluation approach
By day 60
- Own a data product or AI use case
- Set up quality checks and monitoring
- Present findings to stakeholders
By day 90
- Lead delivery of a production AI or data workstream
- Define metrics for success with the client
- Share lessons internally
How we work with clients
- Prepare workshops with a clear decision, relevant evidence, and the people needed to act on the outcome.
- Document assumptions, alternatives, costs, dependencies, and risks so stakeholders can review your recommendation.
- Escalate delivery or quality concerns early with options and a recommended course of action.
- Protect confidential information and use approved collaboration and delivery environments.
- Transfer knowledge through pairing, practical documentation, and walkthroughs with the people taking ownership.
Tools you may use
- Snowflake
- Databricks
- dbt
- Airflow
- Python
- pgvector
- OpenAI / Azure AI
- Power BI
Interview process
- 01
Apply
Share your background. A human reads every application.
- 02
Intro call
30 minutes on your goals, experience and what you want next.
- 03
Technical conversation
A practical discussion with a practitioner — no trick puzzles.
- Technical deep-dive. Discuss a data or AI system you shipped and what you'd change.
- Case discussion. Scope an AI use case: value, risk, evaluation and cost.
- 04
Team & terms
Meet the people you'd work with, then agree scope and terms in writing.
Location, engagement and compensation
This is an expression-of-interest profile for client work we are shaping. A confirmed vacancy, employment arrangement, travel requirements and compensation range have not yet been published. Before an approved role proceeds, we agree client context, availability, duration, compensation or rate and any benefits in writing.
Telugu Technologies is an equal opportunity employer. We consider all qualified applicants regardless of race, color, religion, sex, sexual orientation, gender identity, national origin, age, disability or veteran status, and provide reasonable accommodations on request — email careers@telugutechnologies.com.