Senior AI/ML Engineer (LLM)

@MashoodurRehman1

I am an experienced software engineer with over 5 years in the industry and more than 2 years specializing in building LLM applications in production. I have a strong foundation in Python and have worked extensively with modern LLM tooling. My expertise includes designing complex AI pipelines that transform unstructured data into meaningful insights, aligning with your need to convert engineering scopes into structured construction schedules. My past projects have equipped me with advanced skills in agent frameworks, retrieval-augmented generation, and prompt engineering, all crucial for architecting agentic workflows and retrieval systems in your project. I have engineered evaluation frameworks to ensure AI accuracy, which directly aligns with the need to reduce hallucination and measure schedule accuracy. Collaborating with subject matter experts to encode logic into AI systems is a regular part of my process. I am keen to discuss how my skills can contribute to ScheduleForce's success. Could you share more about the tools currently being used in your pipeline? I am available to provide further details about my experience at your convenience.

@ZohaibRoy

⭐⭐⭐⭐⭐ Build Smart AI Pipelines for Construction Scheduling ❇️ Hi My Friend, I hope you're doing well. I reviewed your project requirements and see you are looking for an AI reasoning expert for ScheduleForce. You don’t need to look any further; Zohaib is here to help you! My team has handled over 50 similar projects, focusing on AI and construction scheduling. I will design efficient workflows and systems to convert data into structured schedules, while ensuring accuracy in every step. ➡️ Why Me? I have over 5 years of experience in software engineering, with 2 years specifically in building LLM applications. My skills include designing AI workflows, prompt engineering, and optimizing system performance. I also have a strong grip on Python and related technologies, which will ensure a smooth execution of your project. ➡️ Let's have a quick chat to discuss your project in detail. I’d love to show you samples of my previous work and explore how we can achieve your goals together. ➡️ Skills & Experience: ✅ AI Workflow Design ✅ Python Programming ✅ LLM Applications ✅ Prompt Engineering ✅ Agent Frameworks ✅ Retrieval Systems ✅ Evaluation Frameworks ✅ Construction Scheduling ✅ Performance Optimization ✅ Data Structuring ✅ Technical Documentation ✅ Non-deterministic Testing Waiting for your response! Best Regards, Zohaib

@gunslinger

Hello, I’ve built production RAG and LLM applications in Python including agentic workflows, retrieval pipelines, tool/function calling and LLM routing. This role is a very strong match for my background because I’ve worked on exactly the problem of turning unstructured information into reliable AI driven workflows, while also putting evaluation and testing around the system instead of trusting the model blindly. I’m comfortable with Python, FastAPI, vector databases/pgvector, PostgreSQL and modern LLM tooling, and I’ve worked with prompt optimization, RAG and multi LLM architectures. For ScheduleForce, I’d focus on making the reasoning pipeline deterministic where it needs to be with proper retrieval, tool boundaries and evaluation around schedule accuracy and hallucinations. I’m also comfortable working directly with domain experts to turn their planning rules into something the system can actually execute. Kindly contct me for further discussion.

@diektech

Hi — Elias here from Miami. I see you're developing an AI-powered construction planning tool aimed at enhancing efficiency and decision-making. What usually matters most here is ensuring the AI integrates smoothly with existing workflows while remaining adaptable for future needs. A common issue in systems like this is balancing the complexity of AI models with user-friendly interfaces. The tricky part is managing data inputs and outputs effectively, especially with large datasets and real-time adjustments. My approach would involve creating a modular architecture, allowing for easy updates and scalability. This ensures the system can evolve as requirements change, focusing on stability and maintainability with robust testing throughout development. I've worked on similar projects involving AI integrations for logistics and resource management, prioritizing user experience alongside technical performance. A few questions to better understand the scope: Q1 – What specific user roles will interact with the system, and what are their key workflows? Q2 – Are there existing data sources or platforms that need to be integrated? Q3 – What are your expectations regarding the AI's decision-making capabilities and level of automation? Happy to discuss the details and suggest the best technical approach. Looking forward to hearing from you.

@Mahihassi

As an AI and Cloud Developer with 5+ years in software engineering, including 2+ years building LLM applications in production, I am well-equipped to own the AI reasoning core of ScheduleForce. I have a strong background in agent frameworks, RAG system development, prompt engineering, and tool/function calling--all of which were prominently mentioned in the job description. These skills will be directly applicable to designing the agentic workflows and retrieval systems required for an LLM to reason about scope, sequencing, and dependencies. Moreover, I have a record of developing evaluation frameworks for non-deterministic AI systems like this one. This experience will be invaluable in measuring schedule accuracy and reducing hallucination – shedding light on what works and optimizing the process further. My ability to translate complex domain logic into reliable AI workflows also aligns perfectly with your role's need to convert engineering scope into specific construction schedules. Lastly, my focus on clean architecture, scalability, and production-ready systems complements the goal of optimizing latency, cost, and reliability of model calls in production. My specialization in building scalable backend systems assures efficiency and robustness in architecting agentic/workflows that you require. I believe my proficiency in Python along with modern LLM tooling signifies a strong fit for this senior role. Let's collaborate on making ScheduleForce a cutting-edge AI project.

@ShaikhAneesa

Hello, I have carefully reviewed the project requirements for the AI-Powered Construction Planning Engineer role. The project involves designing and implementing AI reasoning core for ScheduleForce, focusing on converting engineering scope into construction schedules using LLM technology. Let's chat and discuss it further. To handle your project, I will start with designing and building LLM-powered pipelines to convert engineering scope into schedule activities, durations, and logic. I will then architect agentic workflows and retrieval systems over technical documents, develop evaluation frameworks for schedule accuracy, and collaborate with construction SMEs to encode planning logic. Proficient in Python and experienced in building LLM applications, I am well-equipped to optimize model calls and ensure the reliability of AI workflows. Before signing-off my bid, I would like to ask a question, i.e., how do you envision integrating real-time data updates into the construction scheduling process? Best Regards, Aneesa.

@Herculesz

Hi, this is a real document-reasoning problem, not a simple prompt-wrapping exercise, and I’ve built production AI systems that turn messy source material into structured operational outputs. The real engineering risk here is getting consistent schedule logic from incomplete or ambiguous scope data without letting retrieval gaps or agent drift create plausible but wrong dependencies. The closest matches in my background are AI-Driven Marketing Suite Development -- 2 and Python Bug Localization Using Transformer Models (CodeBERT + TreeBERT). One maps to multi-step AI orchestration over noisy business inputs, and the other maps to evaluation discipline, confidence scoring, and measuring non-deterministic system quality. I usually structure systems like this by separating extraction normalization, planning context retrieval, reasoning/orchestration, and schedule validation into distinct layers. In Python, that keeps the planning core testable and makes it easier to isolate whether failures come from source quality, retrieval, or reasoning. For reliability, I’d recommend grounding checks on every generated activity, confidence thresholds around dependency generation, and an evaluation harness that scores output against SME-reviewed schedules at the rule and graph level, not just text similarity. If useful, I can outline the agent flow and evaluation pipeline for the scheduling core in a short working session. Thanks, Hercules

@kareemmaize

Hi, The key to ScheduleForce is making the LLM reason inside a constrained scheduling system, not letting it invent a plausible-looking schedule and hoping the output is correct. I would structure the pipeline as scope extraction -> work-package decomposition -> targeted retrieval -> activity generation -> dependency construction -> deterministic validation -> repair. The model would handle interpretation and ambiguity, while explicit rules enforce precedence, required activities, valid relationships, duration bounds, and schedule completeness. I would also build evaluation from day one: activity coverage, dependency accuracy, duration error, hallucinated scope, invalid logic, consistency across repeated runs, latency, and model cost, all benchmarked against SME-approved schedules. Relevant work includes a production AI workflow platform with document processing, knowledge retrieval, structured outputs, human review, retries, and auditability, plus Python/FastAPI AI systems using Gemini/Genkit. For production reliability, I prefer typed tool calls, versioned prompts, traceable retrieval, structured outputs, and deterministic validators around the model rather than long free-form agent chains. If it aligns with you, let's discuss in detail via private chat.

@maxscender

This project’s core challenge is converting unstructured scope data into a reliable, logic-driven schedule, where accuracy is the primary metric. My approach centers on enforcing structure: I’ll build an agentic pipeline that uses retrieval to ground every LLM decision in source documents lock-step with tool-use for deterministic calculations, ensuring that sequencing and logic are computed rather than guessed. I will treat schedule generation as a constrained problem, using a validation layer that programmatically checks dependencies and flagging anomalies for SME review, which directly mitigates hallucination. Furthermore, I would implement a scenario-based evaluation framework in which we test against known project parameters to continuously benchmark reasoning accuracy commoditize the feedback loop between the output and the model prompts. This is about converting your current pipeline into a system that verifies its own work while maintaining strict cost and latency controls, rather than relying on freeform generation. The design focuses on creating a reliable, auditable chain from drawing data to deliverable.

@JustinJcob

I got you! I can help build ScheduleForce’s AI reasoning core to turn extracted drawings and scope data into structured schedule activities, durations, sequencing logic, and dependencies. I’m ready to handle the agentic workflows, RAG over technical documents, prompt/tool-calling strategy, evaluation framework, and production optimization. I’m young, a fast learner, and available 24/7 to move fast with your team. Get the demo first before you pay. A couple of quick questions: What format does your extracted drawing/scope data currently come in, and do you already have historical schedules to use as ground truth for evaluation? Also, which LLM stack or agent framework are you currently using, if any? Let’s chat and discuss the answers so I can map the workflow clearly. Kind regards, Haroon Z

@Aaleen110

You need an AI core to turn construction data into a logic-driven schedule, and I will build the LLM reasoning pipeline for that. I will architect agentic workflows using Python, hooking into an LLM to process extracted drawing and scope data, also building retrieval-augmented generation (RAG) systems to pull relevant details from technical documents. The output will be schedule activities, durations, and logic ready for your construction SME to review. I would build the RAG system first so the LLM has access to the necessary context before we define the core reasoning logic. I will develop evaluation frameworks to measure schedule accuracy and reduce hallucination, also designing the AI training data pipeline. Track record on here: 100% on time, 100% on budget, 5.0 across 8 reviews. This shows I can deliver data integration and analysis jobs, like the EUR 3,250 exchange-data integration job, also geotechnical investigation reviews, with consistent quality. How do you want to handle the encoding of planning logic so the AI doesn't just follow a textbook approach but reflects specific company know-how? I need your agreement on the project scope.

@Hubready

Hello!, This is James from Hollywood... I read your AI-Powered Construction Planning Engineer post carefully, and I understand the goal is to build a real AI tool that helps with construction planning, not just a prototype. With 15 years of experience in Python, ML, AI model development, workflow automation, and production software, I can help turn this into something practical, accurate, and scalable. My approach would be: 1. Review your current planning workflow and main pain points 2. Define the best AI use case first, whether that is scheduling support, forecasting, risk detection, or automation 3. Build and test the solution in phases 4. Integrate it cleanly and refine it based on real usage I focus on solutions that are maintainable, useful, and built for production. I also pay close attention to data quality, edge cases, and deployment so the final result works well in the real world. Could you please clarify the following questions to help me better understand the project? 1. What exact planning function should the AI support first? 2. Do you already have historical project data, drawings, or planning docs? 3. Should this be a standalone tool or integrated with an existing system? Relevant work I’ve done includes AI workflow automation for a logistics dashboard, a Python planning assistant for a SaaS ops team, and an internal ML tool for document processing in a construction-related workflow.

@mayr81

With over five years of software engineering experience, with a specific focus on building LLM applications for over two years, I can confidently apply my skills to your project as a Senior AI/ML Engineer for ScheduleForce. My proficiency in Python and modern LLM tooling, combined with my strong experience in agent frameworks, RAG Retrieval systems, and prompt engineering make me equipped to design and build LLM-powered pipelines that turn your extracted drawing and scope data into a structured construction schedule. What sets me apart is my capability to translate complex domain logic into reliable AI workflows - an essential skill for reducing hallucination and ensuring your AI reasoning core functions seamlessly. I have successfully developed evaluation frameworks for non-deterministic AI systems and understand the importance of accuracy and reliability in scheduling. In addition to my technical expertise, I offer clear communication, reliability, efficiency, and a strong collaboration mindset. Your project requires not only technical skills but also coordination across teams. Having worked on different projects involving client communication, reporting, and project coordination, I am confident in my ability to align with Construction SMEs and leverage their expertise to encode planning logic into the system effectively. Let's optimize your construction scheduling processes together!

@sabaehsan

Greetings, Your project focuses on developing an AI-driven system that transforms construction drawings and scope data into a structured schedule. I can help by designing and building pipelines powered by large language models (LLMs) that accurately convert engineering scopes into actionable activities, durations, and logical dependencies. With over five years of software engineering experience, including two years specifically working on LLM applications in production, I have a strong grasp of prompt engineering, agent frameworks, and retrieval systems. My approach would involve collaborating closely with construction subject matter experts to ensure we capture the necessary planning logic, while also implementing frameworks to assess schedule accuracy and minimize errors. I’m excited about the opportunity to contribute to ScheduleForce and help optimize your AI workflows. Best regards, Saba Ehsan

@HamparProd

Hi, ScheduleForce is a strong fit for my work with production LLM systems, RAG, agent/tool workflows and structured AI pipelines. I would approach the reasoning core as a constrained scheduling system rather than asking an LLM to generate a schedule in one pass. Extracted drawing/scope data becomes normalized structured input; retrieval supplies relevant project context and planning rules; specialized agents/tools propose activities, durations and dependencies; then deterministic validation checks sequencing, schema compliance and unsupported assumptions before results are accepted. I can build Python-based RAG and tool-calling pipelines, structured outputs, tracing and evaluation suites that measure schedule accuracy against SME-approved ground truth. I’d also track retrieval quality, hallucinations, dependency errors, latency, token usage and cost so model changes can be tested rather than judged subjectively. I’m comfortable translating evolving domain rules from construction SMEs into reusable tools, constraints and evaluation cases rather than burying them inside prompts. I’m available for a long-term engagement and can start with an audit of the current reasoning pipeline and evaluation baseline.

@starttechno

ScheduleForce needs a dependable reasoning core that transforms extracted drawings and scope data into structured, logic-driven construction schedules—not just generated text. I can design this pipeline in Python using FastAPI, an agent/tool orchestration layer, and a structured schedule schema for activities, durations, predecessors, constraints, and confidence levels. My approach would separate extraction, scope normalization, planning, dependency validation, and schedule assembly into traceable stages. Technical documents can be indexed with metadata-aware RAG, while retrieval tools expose relevant drawing sections, specifications, and construction rules to the planning agent. I’ll use function calling for deterministic outputs and validation tools so the model cannot freely invent activities or logic. SME planning rules can be encoded as explicit constraints and reusable prompt modules rather than hidden assumptions. For reliability, I’ll build an evaluation harness with representative scope/drawing cases, expected activities and relationships, regression tests, structured scoring, and hallucination checks. I’ll also track token usage, latency, retries, and model quality to optimize production calls through caching, model routing, and limited-context retrieval. Muhammad Saad

@vladang

ScheduleForce’s core challenge is turning incomplete scope data into defensible schedule logic—not merely plausible LLM output. Activities, durations, and dependencies need traceable evidence, deterministic validation, and clear handling of uncertainty. I’d model the schedule as a typed dependency graph, with tool calls producing schema-validated activities and retrieval tied to exact document sections. The pipeline would separate extraction, planning, validation, and repair so one weak model response cannot silently corrupt the schedule. Evaluation would compare outputs against SME-reviewed baselines, measuring missing scope, invalid dependencies, duration error, and unsupported assumptions. I’d also test model and prompt changes before release while tracking latency, token cost, and failure rates. The key design choice is the source of truth for planning rules. Are your first evaluation cases based on completed schedules, construction SME annotations, or both?

@jaroslavcaprata

The core challenge here is making construction scheduling deterministic enough to trust while still using LLMs for the parts that genuinely require reasoning over ambiguous scope and technical documents. I would structure the system as a staged pipeline rather than a single large prompt: extract and normalize scope → retrieve relevant planning rules and historical context → generate candidate activities → apply deterministic sequencing/dependency tools → validate the schedule → score and retry questionable outputs. For RAG, I would keep document provenance attached to generated activities so planners can trace why a duration, dependency, or sequencing decision was proposed. Tool/function calling would handle calculations and rule checks that should not be left to free-form model reasoning. I would also build evaluation into the product from the beginning: activity coverage, dependency correctness, unsupported assumptions, schedule consistency, latency, token cost, and comparison against SME-approved schedules. Failed cases can then become regression tests instead of recurring production problems. Python is a strong fit for the orchestration, retrieval, evaluation, and model-service layer. What format does ScheduleForce currently produce after drawing/scope extraction, and do you already have SME-approved historical schedules that can serve as evaluation ground truth?

@abdf2010

Hi, I am a professional web developer and I can do this project "AI-Powered Construction Planning Engineer", I have 5 years of experience in web development. I have done many projects like this. I can do this job for you. I can start right now. Please contact me. Thanks