Description :

Role : ML Engineer

Job Location : Bangalore

Years of Experience : 3 to 5 years of relevant experience

Mandatory Requirement : Passionate about Privacy & Responsible AI

About the Role :

We are building a privacy-first AI platform that secures sensitive data before, during, and after LLM interactions. As an ML Engineer, you will work on core modules from entity recognition and data anonymization to agentic systems and secure LLM orchestration. You will own model selection, training, fine-tuning, and deployment across production pipelines handling structured and unstructured data at scale.

Responsibilities :

- Design, train, fine-tune, and deploy ML models for entity detection, entity masking, text classification, and data anonymization

- Build and optimize NLP pipelines for privacy risk identification and sensitive data tagging across structured and unstructured datasets

- Develop and maintain agentic AI systems that orchestrate multi-step reasoning, tool use, and autonomous task execution

- Implement and evaluate Retrieval-Augmented Generation (RAG) pipelines for secure, context-aware querying

- Work on LLM integration layer : prompt engineering, guardrails, output validation, and data mitigation strategies

- differential privacy and anonymization techniques to ensure data protection before LLM exposure

- Optimize model inference for latency and throughput in production (containerized microservices)

- Collaborate with backend (Go) and platform teams to integrate ML services into existing APIs and Kafka-based event pipelines

- Establish evaluation frameworks : accuracy, bias detection, privacy leakage metrics, and model drift monitoring

- Stay current with state-of-the-art research and recommend adoption of new architectures or techniques.

Must-Have Skills :

Development :

- Python Expertise : Professional experience building and scaling Microservices.

- Frameworks : Proficiency in FastAPI (for high-perf APIs) and Flask.

- Concurrency : Deep comfort with Asynchronous development (asyncio).

- Distributed Systems : Hands-on experience with Redis and Celery for background task orchestration.

LLMs & Applied NLP :

- Strong understanding of Encoder models (BERT, RoBERTa, DeBERTa)

- Strong understanding of Decoder models (GPT family, LLaMA, Mistral)

- Deep knowledge of Transformer architecture attention mechanisms, positional encoding, k-v caching, tokenization strategies

- Familiarity with Dense, Convolutional, Sequence model layers (LSTMs, GRUs)

- Hands-on experience with LLM fine-tuning : LoRA, QLoRA, PEFT, instruction tuning

- Prompt engineering : few-shot, chain-of-thought, structured output enforcement

- Named Entity Recognition (NER) : custom entity models for PII (names, addresses, financial data, health records)

- Text classification, sentiment analysis, and zero-shot/few-shot classification (e.g., DistilBART-MNLI)

- Familiarity with embedding models and vector similarity search (FAISS, Pinecone, ChromaDB)

Agents & Orchestration :

- Experience building AI Agents : tool-calling agents, ReAct pattern, plan-and-execute architectures

- Familiarity with agent frameworks : LangChain, LangGraph, or equivalent

- Understanding of multi-agent systems : agent communication, task delegation, state management

- RAG pipeline design : chunking strategies, retrieval ranking, context window management

Qualifications :

- B.Tech/M.Tech/MS in Computer Science, AI/ML, Mathematics, or related field

- 3-5 years of professional experience building and deploying ML models in production

- Published or demonstrable work with transformer-based architectures

- Strong fundamentals in linear algebra, probability, statistics, and optimization.

Why Join Us ?

- Work at the intersection of AI, privacy, and compliance, shaping global standards.

- Be part of an award-winning startup making an impact worldwide.

- Take ownership and drive innovation in Responsible AI & Privacy Tech.

About PrivaSapien :

Privacy Engineering and Responsible AI are critical for building ethical AI guardrails. At PrivaSapien, we are pioneering a comprehensive suite of Privacy Engineering and Responsible AI products to govern data and AI the right way.

Recognized globally, we have won multiple awards, including :

- Morgan Stanleys "Future of AI" Award

- Accenture Ventures "AI : Digital Core" Award

- Google for Startups AI First Award

- NASSCOM Emerge 50 Award

- NITI Aayogs Aatmanirbhar Bharat Award

- DSCIs Innovation Box Award

- SDAIA PET Sandbox Graduation (Saudi Arabia)

- UK Gov & Infosys Fairness Award

We have secured funding from a US-based VC firm and recently turned down an acquisition offer from a $Trillion companya testament to our vision and impact.

We serve large enterprises across India, the Middle East, Europe, and the US and have filed 25 patents. Our team is deeply research-driven, with 40% of our talent recruited from IISc & ISI Kolkata.

Join us in shaping the future of AI with privacy at its core Our Vision & Work Culture :

1. Vision : Leading the Evolution of the Privacy Era

At PrivaSapien, we believe privacy is a fundamental right in the digital age. Without it, human thoughts, emotions, and even decisions can be manipulatedturning individuals into "AlgoSlaves." Our mission is to build privacy engineering & Responsible AI solutions that safeguard digital experiences while creating strategic advantages for businesses.

2. Passion : Crazy Enough to Change the World

We are a team of privacy and Responsible AI enthusiasts who believe in our power to protect every individual's privacy globally. Our relentless pursuit of innovation fuels our mission.

3. Integrity & Trust : The Foundation of Privacy

Integrity is non-negotiable at PrivaSapien. In the world of privacy, trust is everythingour customers rely on us to safeguard their data, and we uphold that responsibility with the highest ethical standards.

4. Customer First, Customer Last

We are obsessed with customer value. Every decisionbig or smallis driven by our customers' needs, expectations, and preferences.

5. Adaptability : Thriving in Change

As Darwin said, It is not the strongest, but the most adaptable that survives. In an era of rapid innovation, we embrace change, agility, and continuous learning to stay ahead.

6. Innovation : Setting the Industry Standard

We dont follow trends; we create them. Every PrivaSapien is empowered to innovate in their role, redefining privacy and Responsible AI for the world.

7. Perseverance : Privacy is Hard, But Worth It

Protecting privacy is uncharted territorya difficult yet essential mission. We are prepared to push boundaries, embrace challenges, and work tirelessly to show the world the right path to privacy.

8. Experimentation : Fail Fast, Learn Faster

We encourage structured rapid experimentationfailing quickly, learning effectively, and making strategic, high-impact decisions that drive industry leadership.

9. Collaboration : One Team, One Mission

We are a close-knit family, united by a shared vision. Supporting and uplifting each other in tough times makes us stronger. Together, we will build a safer digital world.

10. Fun : Work Hard, Laugh Harder

Life is too short to not enjoy the journey. We foster a positive, transparent, non-hierarchical environment where passion, creativity, and fun thrive.