Location: Bangalore

Type: Full-time

Who We Are

KodNest is an EdTech headquartered in Bangalore, redefining how students transition from “just graduates” to job-ready professionals. We believe in transparency, accountability, and continuous learning.

But we’re not stopping at being average edtech or typical “view your grade” setups. KodNest is on a mission to become an AI-native EdTech platform, delivering dynamic, data-driven micro-learning experiences.

Role Summary

We’re seeking a Full-Stack AI/Data Engineer to design and implement complex ML/DL solutions from the ground up. You’ll train or fine-tune models (including large language models, advanced neural nets), develop sophisticated algorithms to meet defined outcomes, and integrate them into our product’s data pipelines and user-facing features. This isn’t a simple “plug in an external API” role—it requires hands-on model building, algorithm design, and full-stack integration skills.

Key Responsibilities

• Model Development & Training
• Design and build advanced machine learning or deep learning models (NLP, LLMs, recommendation systems, etc.) from scratch or by fine-tuning existing architectures.
• Oversee data preparation (data cleaning, labeling, feature engineering) and model training (hyperparameter tuning, optimization).
• Implement model evaluation strategies (cross-validation, A/B testing) to ensure reliable performance against defined KPIs.

Algorithm Design & Custom Logic

• Craft bespoke algorithms or similar approach to provide the precise responses or outcomes required—e.g., personal micro-coaching, advanced readiness scoring, custom Q&A flows, or domain-specific code analysis.
• Integrate novel heuristic approaches or multi-step pipelines if needed, blending rule-based and learned components.
• Collaborate with product or domain experts to ensure the algorithmic logic aligns with real-world constraints and user expectations.

Full-Stack Data & Pipeline Ownership

• Build end-to-end solutions: from ingesting raw logs/events → data preprocessing → model training → deployment → real-time inference.
• Manage infrastructure on AWS/GCP/Azure (serverless or container-based) for both batch/offline training jobs and production inference.
• Ensure robust and scalable pipelines that handle large volumes of data with minimal latency and high reliability.

Front-End & API Integration

• Expose the trained models or algorithms via APIs that feed into the front-end UI.
• Collaborate with front-end engineers or handle it yourself (if full-stack) to surface model outputs (e.g., readiness scores, recommendations, Q&A) in a clean, intuitive manner.
• Provide dynamic, real-time updates as the system ingests new data or user events.

Deployment, Monitoring & Optimization

• Containerize and deploy ML services (Docker/Kubernetes/ECS), implementing CI/CD for continuous iteration.
• Set up monitoring for model performance (accuracy drift, latency, resource usage).
• Optimize cost and scale (autoscaling GPU/CPU resources) while retaining stable user-facing performance.
• Research & Innovation
• Stay abreast of state-of-the-art in neural networks, LLM fine-tuning techniques, advanced recommender systems, etc.
• Experiment with new architectures (transformers, RL, generative models) that could yield better outcomes.
• Potentially publish or share internal research on the custom methods you develop, if relevant to the product roadmap.
• Collaboration & Stakeholder Engagement
• Work closely with data scientists, domain experts, or product leads to refine model goals, interpret results, and iterate on improvements.
• Communicate complex ML concepts to non-technical audiences—explaining why a certain model or algorithm approach was chosen.
• Engage in agile product cycles: gather user feedback, adjust model strategies, and deliver incremental improvements rapidly.

Skills & Qualifications

Technical Must-Haves

Advanced ML/DL Proficiency

Strong track record in training models from scratch using frameworks like PyTorch or TensorFlow.

Ability to fine-tune pre-trained LLMs (GPT, BERT, T5, or custom domain-specific models) or other advanced nets.

Solid theoretical grounding in neural network architectures, optimization methods, overfitting controls, etc.

Algorithm Design & Data Structures

Proven experience devising custom algorithms for domain-specific needs—beyond typical scikit-learn out-of-the-box solutions.

Comfortable blending multiple approaches: rule-based, heuristic, learned models, and ensemble strategies.

Skilled in analyzing large, potentially messy datasets for feature extraction or labeling.

Full-Stack Data Engineering

Proficiency in backend (Node.js, Python, Go, or similar) and data pipeline orchestration (Airflow, Step Functions, serverless frameworks).

Familiarity with distributed computing (Spark, Ray) or HPC for heavy training jobs is a plus.

Knowledge of relevant cloud services (AWS S3, EC2, ECS/EKS, GCP equivalents) for large-scale training, inference, data storage.

Front-End Integration (Nice to have)

Comfortable exposing model outputs via APIs and, if needed, building web components (React/Angular/Vue) to display real-time results or interactive AI-driven features.

Solid grasp of user experience concerns—making complex AI outputs explainable, with reason codes or confidence levels, etc.

DevOps & MLOps

Experience with containerization (Docker), orchestration (Kubernetes/ECS), continuous integration/deployment.

Familiarity with model serving solutions (TorchServe, TF Serving, custom microservices) and monitoring tools.

Ability to optimize GPU usage, set up HPC clusters, or cloud-based ML pipelines as required.

Soft Skills

Analytical Problem-Solving: Approach ambiguous tasks with structured experimentation, pivoting quickly based on results.

Collaboration: Willingness to brainstorm with product leads or subject matter experts to shape ML goals.

Communication: Translate complex ML constraints or results into accessible insights for stakeholders.

Ownership: Drive projects from concept to production, and continuously refine them based on user feedback or performance metrics.

Ideal Candidate Profile

5+ years in data engineering or machine learning roles, with at least 2-3 real-world deep learning projects deployed in production.

Expertise in fine-tuning modern LLMs or large neural networks for specialized tasks (NLP, recommendations, complex classification).

Comfortable building entire data and ML lifecycle (ingestion → training → inference → monitoring).

Thrives in a fast-paced environment, shipping iterative improvements, balancing research with pragmatic deliverables.

Passion for user impact: sees model building as a means to deliver better experiences or outcomes, not just a research exercise.

Why This Role Matters

We are on the leading edge of delivering advanced AI-driven features—ranging from deep analysis of user behavior to custom model responses that guide user actions. This position is our key AI expert role, shaping how we design, train, and deploy the complex models that set our product apart.