Ai/ml Engineer Job in Roxiler Systems

AI/ML Engineer

Experience: 2 3 Years
Education: BBA, BSc, BTech/BE, MSc, MTech, MCA
Location: Pune

About the Role

We are seeking a passionate AI/ML Engineer to design, build, and scale production-grade AI systems that power our platform. You will work on real-world AI problems such as document intelligence, real-time voice AI, and intelligent job-to-candidate matching systems. This role requires strong hands-on experience in LLM engineering, backend development, and AI system productionization.

Required Skills

• Python (FastAPI)
• LangChain, LlamaIndex
• OpenAI API, Anthropic

Good to Have

• Postgres (Pgvector), Supabase
• Redis
• Docker
• LangSmith
• GitHub Actions

Roles & Responsibilities

• Build Agentic Workflows: Design and deploy AI agents capable of reasoning and decision-making using LangChain or LlamaIndex.
• Develop Voice AI Systems: Build low-latency, real-time conversational voice bots for candidate outreach using WebSockets, STT, and TTS, ensuring effective context retention and state management.
• Engineer Data Pipelines: Create robust parsing systems that enforce strict JSON outputs from LLMs for resume data extraction, along with data cleaning pipelines for unstructured formats (PDF/DOCX).
• Implement Advanced RAG Systems: Develop retrieval pipelines using Pgvector or Pinecone with hybrid search (semantic + keyword) for accurate job-to-candidate matching.
• Productionize & Monitor AI Systems: Set up tracing and observability using LangSmith to debug chains, monitor token usage, and optimize performance and cost.
• Backend Integration: Package AI logic into scalable, asynchronous microservices using FastAPI and Docker.

Technical Requirements

1. Generative AI & LLM Engineering

• Experience enforcing structured JSON outputs using function calling.
• Strong prompt engineering skills (Chain-of-Thought, Few-Shot).
• Hands-on experience orchestrating complex AI pipelines using LangChain or LlamaIndex.

2. Voice AI & Real-Time Processing

• Experience with STT/TTS APIs (e.g., Whisper, Deepgram).
• Strong understanding of WebSockets and async programming for real-time audio streaming.
• Ability to design conversation managers with memory and context handling.

3. Search & Data (RAG)

• Experience with vector databases such as Pgvector, Pinecone, or Qdrant.
• Expertise in chunking, cleaning, and ingesting unstructured data.

4. MLOps & Production Engineering

• Observability and tracing using LangSmith.
• Writing automated AI evaluation scripts to test prompts before deployment.
• Experience optimizing cost by balancing model performance and token usage.

5. Core Backend Engineering

• Strong Python skills with FastAPI.
• Proficiency in async/await patterns for concurrent processing.
• Experience with Docker and relational databases (SQL).

6. Machine Learning & Algorithms

• Understanding of recommendation systems (Collaborative Filtering, Two-Tower models).
• Experience with ranking and re-ranking techniques (LTR, Cross-Encoders).
• Familiarity with traditional ML models using Scikit-learn or XGBoost.

The Applied Mindset We Value

• Model Strategy: Ability to choose the right model based on cost, latency, and intelligence requirements.
• Security First: Awareness of prompt injection and jailbreaking risks and strategies to mitigate them.
• Hallucination Control: Strong grounding techniques to ensure factual and reliable AI outputs.

Tech Stack Overview

• Languages: Python (FastAPI)
• AI & Orchestration: LangChain, LlamaIndex, OpenAI API, Anthropic
• Voice AI: Deepgram, Twilio
• Databases: Postgres (Pgvector), Supabase, Redis
• DevOps & MLOps: Docker, LangSmith, GitHub Actions