Junior AI ML programmer

Amplelogic; Hyderabad; On-site; INR 600,000 - 900,000; Full time

Job summary

Amplelogic in Hyderabad is seeking a Junior AI/ML Programmer to help build data pipelines, train models, and contribute to RAG/CAG features on our on-prem AI/ML platform, guided by senior architects and engineers.

You will deploy containers, optimize GPU usage, and collaborate on cloud deployments when required, while documenting results and ensuring compliance with strict data-privacy standards.

Qualifications

• B.Tech/BE or equivalent in CS with strong academic record.
• Completed AI/ML training programs with certification.
• Internship in DS/DE/ML Ops with outcomes.
• Strong Python and ML knowledge.
• Excellent written and verbal English communication.

Responsibilities

• Build, clean, and validate datasets and data pipelines.
• Train, fine-tune, and evaluate ML models and select appropriate classes.
• Implement LLM and agentic AI components (RAG/CAG, tools, multi-step agents).
• Package and deploy models on on-prem GPU infrastructure (Docker, serving, monitoring) and cloud as needed.
• Write documentation, evaluation reports, and test cases; present results to stakeholders.
• Follow quality, security, and data-privacy practices for regulated software.

Skills

Python scripting

Data pipelines

ML model selection

English communication

Git

Education

B.Tech/BE or MSc CS

Tools

Docker

NumPy/Pandas

Job description

Junior AI/ML Programmer (5 openings)

Location: Hyderabad (on-site) | Team: AI/ML Platform | Level: Junior (0-2 years post-degree)

About the role

We build an on-premises AI/ML platform that powers products used by pharmaceutical and life-sciences companies worldwide. Our models and pipelines run inside customer-controlled, GPU-equipped environments under strict quality and data-privacy rules (GxP, 21 CFR Part 11, HIPAA/GDPR). As a Junior AI/ML Programmer, you will build and ship real components of this platform - data pipelines, model training and evaluation code, RAG and agentic AI application features, and deployment/monitoring tooling - under the guidance of senior architects and engineers.

What you will do

• Build, clean, and validate datasets; write robust Python data pipelines.
• Train, fine-tune, and evaluate ML models (classical ML and deep learning) and select the right model class for each data problem.
• Implement LLM and agentic AI application components: retrieval-augmented generation (RAG) and cache-augmented generation (CAG), agentic workflows (tool use, multi-step agents), prompt pipelines, and evaluation harnesses.
• Package and deploy models on our on-prem GPU infrastructure (Docker, model serving, monitoring) and on cloud where required (AWS/Azure/GCP).
• Write clear documentation, evaluation reports, and test cases; present results to technical and non-technical stakeholders.
• Follow our quality, security, and data-privacy practices for regulated software.

Required qualifications

• B.Tech/BE, BCA, B.Sc (CS), MCA, or M.Sc. (CS) with a strong academic record.
• Flawless verbal and written communication and articulation in English (tested directly in interviews).
• Completed AI/ML training program(s) with certification in at least a few areas (e.g., NVIDIA DLI/NCA, DeepLearning.AI, AWS/Azure/GCP ML certifications, university-affiliated programs).
• Strong, verifiable public profiles: LeetCode and/or HackerRank, Kaggle and/or DrivenData, Hugging Face, and GitHub. Links are mandatory in the application; profiles are verified live in interviews.
• At least one (preferably 2-3) internship in Data Science, Data Engineering, or AI/ML Ops, with referenceable outcomes.
• Expert-level Python scripting (idiomatic Python, NumPy/pandas, debugging, testing) and sound knowledge of ML model families and selection criteria for

different data problems (tabular, text, image, time series; small-data vs. big-data regimes).

• Working knowledge of AI/ML infrastructure and hardware (GPU vs. CPU, memory sizing, containers) and at least one cloud (AWS, Azure, or GCP).

Preferred

• Exposure to LLM and agentic AI tooling (Hugging Face Transformers, LangChain/LangGraph, CrewAI/Autogen, MCP, vector databases), PEFT/LoRA fine-tuning, awareness of CAG and long-context techniques, and evaluation frameworks.
• Kaggle/DrivenData competition results with honest write-ups; open-source contributions.
• Awareness of regulated-industry software practices (validation, audit trails, data privacy).