L

Full Stack Development - Internship (WFH)

Lacritz.ai Verified employer

Actively hiring · Posted 6d ago · JEV261113B809C3A0

Salary₹1,32,000 – ₹1,62,000 per year
ExperienceFreshers welcome
LocationBengaluru, Karnataka
Work modeRemote · Internship
Openings1
IndustryComputer Services

Job description

About the internship:Key responsibilities:

1) Local LLM evaluation: Evaluate open-weight models for answer quality, response speed, memory usage, and hardware requirements using tools such as llama.cpp or Ollama.

2) Dataset preparation and fine-tuning: Prepare appropriately sourced training data and fine-tune selected models using LoRA/QLoRA, Hugging Face Transformers, and PEFT. Maintain separate evaluation data and compare results against the base model.

3) Inference optimization: Explore quantization, GPU acceleration, context management, and caching. Stream generated answers into the interface and measure time to first useful answer, generation speed, and resource usage.

4) Audio and transcription: Integrate microphone/system audio capture and local speech-to-text. Handle partial transcripts, pauses, background noise, and different accents.

5) Desktop development: Help build the Windows and macOS application using Electron or a comparable framework. Connect the interface to local AI services and support model downloads, loading states, error handling, and packaging.

6) Personalization: Process uploaded resumes, notes, and reference documents. Use retrieval and conversation context to produce relevant answers and reduce unsupported responses.

7) Testing and documentation: Evaluate end-to-end latency, answer quality, and stability on representative hardware. Document experiments, setup instructions, limitations, and technical blockers.

Required experience:

A working AI application you can demonstrate and explain.

Practical experience running an LLM locally, beyond calling cloud APIs.

Strong Python programming and debugging skills.

Familiarity with PyTorch or Hugging Face Transformers.

Understanding of model inference, context windows, and hardware constraints.

Git and clear communication.

Strongly preferred:

A completed LoRA/QLoRA fine-tuning experiment.

Experience with quantization, GGUF, llama.cpp, or Ollama.

Speech-to-text, streaming audio, or voice assistant projects.

JavaScript/TypeScript, React, and Electron.

Document retrieval/RAG and Windows or Apple Silicon development.

You do not need experience with every tool listed. Practical project ownership and problem-solving matter most.

Initial milestone:

Build a prototype on one operating system that captures and transcribes audio locally, identifies questions, and streams answers from a local LLM. Record latency, answer quality, and memory usage on documented hardware. Evaluate fine-tuning improvements before expanding to the second operating system.

To apply, share a demo or project walkthrough, a GitHub link or code sample if available, and a brief explanation of your contribution. Include any local inference or fine-tuning results, the hardware used, and known limitations.

Shortlisted candidates may be invited to a project walkthrough and a small, paid practical assessment with an agreed scope.Who can apply:Only those candidates can apply who:are available for the work from home job/internshipcan start the work from home job/internship between 28th Oct'26 and 2nd Dec'26are available for duration of 6 monthshave relevant skills and interests* Women wanting to start/restart their career can also apply.are Computer Science Engineering studentsStipend:INR₹ 11,000 - 13,500 /monthDeadline:[hidden]:59:59Other perks:Certificate, Letter of recommendationSkills required:Java, JavaScript, Python, Git, Machine Learning, React, REST API, iOS, Android, OpenAI API, Machine Learning Operations (MLOps), ChatGPT, Claude, WebSockets and WindowsOther Requirements:Candidates who have built at least one working AI application through a personal project, hackathon, freelance assignment, or previous internship are preferred. Please share a demo, GitHub link, or short project walkthrough.Hands-on experience integrating LLM APIs into an application is preferred. Experience with speech-to-text, voice assistants, streaming responses, or document-based question answering is a strong plus.Comfortable building with Python and/or JavaScript/TypeScript. Familiarity with React, backend APIs, and WebSockets is desirable.Able to explain their own code, troubleshoot issues, and take ownership of clearly defined features with regular feedback.Students and recent graduates with strong practical projects are encouraged to apply; prior professional experience is not required.About Company:Building an AI agentic platform for digital marketing teams so that we can automate the entire digital marketing. Additional information:

Stipend Structure:

Fixed pay: ₹ 10,000 - 12,000 /month

Incentive pay: ₹ 1,000 - 1,500 /month

Responsibilities

Key responsibilities:

1) Local LLM evaluation: Evaluate open-weight models for answer quality, response speed, memory usage, and hardware requirements using tools such as llama.cpp or Ollama.

2) Dataset preparation and fine-tuning: Prepare appropriately sourced training data and fine-tune selected models using LoRA/QLoRA, Hugging Face Transformers, and PEFT. Maintain separate evaluation data and compare results against the base model.

3) Inference optimization: Explore quantization, GPU acceleration, context management, and caching. Stream generated answers into the interface and measure time to first useful answer, generation speed, and resource usage.

4) Audio and transcription: Integrate microphone/system audio capture and local speech-to-text. Handle partial transcripts, pauses, background noise, and different accents.

5) Desktop development: Help build the Windows and macOS application using Electron or a comparable framework. Connect the interface to local AI services and support model downloads, loading states, error handling, and packaging.

6) Personalization: Process uploaded resumes, notes, and reference documents. Use retrieval and conversation context to produce relevant answers and reduce unsupported responses.

7) Testing and documentation: Evaluate end-to-end latency, answer quality, and stability on representative hardware. Document experiments, setup instructions, limitations, and technical blockers.

Required experience:

A working AI application you can demonstrate and explain.

Practical experience running an LLM locally, beyond calling cloud APIs.

Strong Python programming and debugging skills.

Familiarity with PyTorch or Hugging Face Transformers.

Understanding of model inference, context windows, and hardware constraints.

Git and clear communication.

Strongly preferred:

A completed LoRA/QLoRA fine-tuning experiment.

Experience with quantization, GGUF, llama.cpp, or Ollama.

Speech-to-text, streaming audio, or voice assistant projects.

JavaScript/TypeScript, React, and Electron.

Document retrieval/RAG and Windows or Apple Silicon development.

You do not need experience with every tool listed. Practical project ownership and problem-solving matter most.

Initial milestone:

Build a prototype on one operating system that captures and transcribes audio locally, identifies questions, and streams answers from a local LLM. Record latency, answer quality, and memory usage on documented hardware. Evaluate fine-tuning improvements before expanding to the second operating system.

To apply, share a demo or project walkthrough, a GitHub link or code sample if available, and a brief explanation of your contribution. Include any local inference or fine-tuning results, the hardware used, and known limitations.

Shortlisted candidates may be invited to a project walkthrough and a small, paid practical assessment with an agreed scope.

Key skills

How to apply

  1. Sign in or create your free profile — takes about two minutes.
  2. Apply with one tap — your profile is shared directly with Lacritz.ai.
  3. Track your application and get updates by email and WhatsApp.
Apply now
You may also like

Similar jobs

More jobs
₹2,00,000 – ₹3,50,000 /year 1–4 yrs Bhiwandi, Dombivali, Kalyan, Thane, Maharashtra Remote · Full time
inventory managementwarehouse operationserpstock audit
Today · JEV2611A69D92EBB9View & apply
Full Stack Development - Internship (WFH)₹1,32,000 – ₹1,62,000 /year · Bengaluru
Apply